Public Response to the Constitutional Court’s Decision on Indonesia’s 2024 Elections

Public Response to the Constitutional Court’s Decision on Indonesia’s 2024 Elections | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Public Response to the Constitutional Court’s Decision on Indonesia’s 2024 Elections Indra Surya Permana, Fardhoni Fardhoni, Christina Juliane This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4482093/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This research investigates the sentiment analysis of public reactions on Twitter to the Constitutional Court’s decision regarding the 2024 Indonesian election. The study focuses on evaluating the effectiveness of Naive Bayes and Gradient Boosted Machines (GBM) in categorizing Twitter sentiments into positive, negative, or neutral. Utilizing TF-IDF vectorization to process the data, our analysis aimed to discern which model more accurately captures the nuances of public sentiment. The results indicate that while Naive Bayes shows high precision and recall in detecting positive sentiments, it performs less effectively for negative and neutral sentiments. In contrast, GBM offers a more uniform performance across all sentiment categories, with particularly strong detection capabilities for neutral sentiments. This comparative analysis underscores the strengths and limitations of each model, providing valuable insights for selecting appropriate sentiment analysis tools depending on the specific nature of the sentiment being analyzed. This study contributes to the strategic application of sentiment analysis models in monitoring and interpreting public opinions in politically significant contexts. Sentiment Analysis Naive Bayes Gradient Boosted Machines Twitter Indonesian Election Figures Figure 1 I. INTRODUCTION The 2024 General Election in Indonesia is an important event that attracts great attention from the public and is in the spotlight. In this context, several important dynamics have already occurred. First, technology, especially social media, has become the main platform for presidential candidates and their supporters to convey their vision, mission, and work programs [ 1 ]. Second, political education and understanding the impact of political decisions are becoming increasingly important in the context of voter participation. Third, global issues such as climate change, energy security, and global health are the main focus in candidate debates [ 2 ]. Fourth, the election, especially the 2024 Presidential Election, reflects the spirit of inclusivity with many candidates from various backgrounds and community groups. Fifth, political polarization, potential abuse of authority, and the threat of fertile dynastic politics are issues that color the course of the 2024 presidential election. Sixth, young people aged 22–30 years will dominate the electorate nationwide, with a share of 56%, or about 114 million. Seventh, misinformation and disinformation on social media are still major issues ahead of the 2024 elections [ 3 ], [ 4 ]. The 2024 election has been completed when the results of the recapitulation are announced in March 2024, where the presidential candidate and vice president pair number 2 get 96,214,691 votes or more than 57% of the total votes and win a victory that is far adrift from the pair number 1 and 2. however, the results of the 2024 election in Indonesia have caused various controversies and debates, including the existence of 718 disputes recorded in the 2024 Presidential & Vice Presidential Election Dispute List [ 5 ]. The Constitutional Court (MK) decided to reject all claims over the results of the 2024 Presidential Election. However, many parties felt disappointed and doubted the decision. Currently, platform X or Twitter is often considered to be a source of information in the analysis of public opinion [ 6 ]. This is because Twitter is a place for people to share opinions, express emotional conditions such as happiness, disappointment, anger, and so on, including political views of a person or group [ 7 ], [ 8 ]. The opinions of social media user X (Twitter) regarding the results of the Constitutional Court (MK) decision on the 2024 election are very diverse and reflect the existing political dynamics. Some Twitter users showed their support for the Constitutional Court's decision, while others expressed disappointment and doubt. There were also a number of users who questioned the process and legal basis of the decision. For example, some have questioned the independence of the election administration agency and highlighted allegations of abuse of power by the government. The behavior of social media users is discussed in research conducted by Septian, et al [ 9 ] and Vindua and Zailani [ 8 ] which discusses the sentiment of Twitter users towards a legal ruling in Indonesia on a national scale. In the digital age, the abundance of sentiment data from social media and online platforms provides a valuable source of information to analyze public views to see the sentiments or opinions formed. However, the complexity and volume of this data demands a sophisticated approach to its analysis. This is where Naive Bayes models and Gradient Boosted Machines play an important role in opinion mining. Naive Bayes, with its prominence in text classification, has become one of the popular approaches in sentiment analysis as has previous research conducted by Isnain, et al [ 10 ], Zahra [ 11 ], and Ansari [ 12 ]. While Gradient Boosted Machines (GBM) models are powerful methods in machine learning that can produce accurate prediction models [ 13 ], [ 14 ]. In the context of sentiment analysis, gradient boosted machines can be used to model non-linear relationships between features in text and existing sentiment[ 15 ], [ 16 ]. Using these models, opinion mining or sentiment analysis of the Constitutional Court's decision related to the 2024 election can be carried out to identify patterns and trends in public opinion. The sentiment analysis can provide an idea of how the public responded to the Constitutional Court's decision, whether the majority supported or questioned the decision, and what factors influenced their views. Research Objectives To find out User X's (Twitter) Opinion Mining on the Constitutional Court Ruling on the 2024 Election with Naive Bayes and Gradient Boosted Machines models. II. LITERATURE REVIEW Text Analysis in Natural Language Processing (NLP) is one of the methods in machine learning (machine learning) used to understand and extract information from text [ 17 ], [ 18 ]. This method involves processing and reading text using computer algorithms to recognize patterns and look up the meaning of the text. These methods of text analysis can include identification of entities such as names of people or places, classification of text into specific categories, information extraction, sentiment analysis, and others. Opion mining or often called Sentiment analysis is a method used in text analysis to identify, extract, and analyze sentiment or opinion. Opion mining is important because with the increasing complexity and volume of data generated by social media and online platforms, it is necessary to understand public opinion quickly and accurately with the help of machine learning [ 19 ], [ 20 ]. Some methods in machine learning go into simple methods and also deep learning or deep learning. One method in opinion mining that is often done is to use the Naïve Bayes Classifier (NBC) method. The NBC method itself is a simple and efficient method that can be used to classify text into different sentiment categories [ 10 ], [ 21 ]. In the Naive Bayes method, probabilistics are used to classify text into sentiment categories based on the likelihood of a sentiment occurring based on the linguistic features present in the text. Gradient Boosted Machines (GBM) are machine learning techniques that use a weak set of predictive models (usually decision trees) to build powerful predictive models incrementally [ 22 ]. The algorithm improves prediction by iteratively adjusting a new model that focuses on cases incorrectly predicted by previous models in the series. Gradient Boosted Machines (GBM) models are sophisticated machine learning techniques for regression and classification. It belongs to the category of 'ensemble learning' algorithms, where multiple predictive models are combined to improve the accuracy and power of predictions [ 23 ]. Both NBC and GBM have distinct roles in sentiment analysis. While Naïve Bayes offers speed and simplicity, making it suitable for initial screenings of sentiment in large datasets, GBM provides depth and precision, particularly valuable in contexts where the differentiation of subtle sentiment nuances is critical. Together, these models contribute to a more holistic sentiment analysis process, allowing for both broad and deep analyses of public opinions on platforms like Twitter regarding critical events such as the Constitutional Court decisions i.e regarding Indonesian 2024 electoral decision [ 10 ], [ 22 ]. The relevance of these models extends to analyzing public reactions to significant societal events, such as judicial decisions. By applying these sentiment analysis models, researchers can capture a broad spectrum of public opinions, providing valuable insights into societal perceptions and influencing future policy formulations [ 19 ], [ 20 ]. This is particularly pertinent in the context of electoral decisions or constitutional judgments, where public sentiment can significantly impact the sociopolitical landscape. The following “State of the Art” table synthesizes significant studies in this area, summarizing their methodologies, key findings, and contributions to the discipline. The table aims to highlight how various approaches have been utilized to enhance our understanding of public opinion and sentiment, reflecting the diverse applications of these technologies across different research contexts Table 1 State of The Art Study Methodology Key Findings Contribution [ 1 ], [ 6 ] Social media monitoring, AI-based Twitter analysis Showed social media's role in election oversight and developed AI frameworks for evaluating government campaign strategies. Highlighted social media and AI's roles in political monitoring and strategy assessment. [ 3 ], [ 4 ] Descriptive analysis, Socio-political studies Described public reactions to judicial decisions and examined electoral conflicts and polarization. Offered insights into public responses and political dynamics surrounding elections. [ 7 ], [ 8 ], [ 9 ] Sentiment analysis with Twitter data, IndoBERT application Applied sentiment analysis to predict election results and analyze legal sentiments. Showcased sentiment analysis's utility in predicting political outcomes and legal reactions. [ 10 ], [ 11 ], [ 21 ] Naïve Bayes for sentiment analysis on government policies and education Evaluated public sentiments on policies and educational initiatives, demonstrating broad effectiveness. Enhanced understanding of public responses to governmental and educational sectors through NLP. [ 22 ], [ 23 ] GBM and ensemble learning Discussed GBM's role in complex classifications and its use to boost prediction accuracy. Demonstrated GBM's capability in detailed sentiment analysis and improved prediction reliability. [ 19 ], [ 20 ] Machine learning in opinion mining, enriched joint sentiment-topic models Investigated opinion mining challenges and introduced new joint sentiment-topic models. Introduced innovative methods for detailed opinion analysis from vast datasets. III. METHODS This research employs opinion mining techniques using the Naïve Bayes Classifier (NBC) and Gradient Boosted Machines (GBM) to analyze sentiments expressed on Twitter regarding the Constitutional Court’s decision on the 2024 Indonesian Presidential Election. The NBC method is a probabilistic classifier known for its simplicity and effectiveness in text classification, which categorizes text based on the likelihood of sentiment occurrence, considering the linguistic features present in the text. Concurrently, GBM enhances prediction accuracy by iteratively refining models that focus on misclassified cases by previous iterations. The following are the stages that will be carried out in sentiment analysis in this study: 1. Data Collection The process of collecting data relevant to the research. In this context, it involves collecting tweets or text from platform X related to the research topic. 2. Dataset The dataset used is a collection of Platform X (or Twitter) comments that use hashtags related to the Constitutional Court's decision trending 5 days before the day when the constitutional court's final decision regarding the Dispute over the Results of the 2024 Presidential and Vice Presidential Elections on April 22, 2024. The dataset is all trending topics with queries BarengPrabowoBangunBangsa, Gibran, Constitutional Court, Constitutional Court Verdict, PeopleWin, Anies Muhaimin, and Ganjar Mahfud. The total data that can be retrieved is 11,721 lines that include username, tweet ID, date, Tweet text content, and source. 3. Data Pre-processing The stage of preparing the raw text for analysis by cleaning it of unnecessary elements or not providing important information. These include: 1. Case Folding: Normalizing all text to lowercase to ensure uniformity. 2. Filtering: Removing non-essential characters like numerals and symbols to focus purely on textual data. 3. Stopword Removal: Eliminating common words that offer minimal unique information for analysis. 4. Tokenizing: Segmenting text into individual words or tokens. 5. Stemming: Reducing words to their base or root form to simplify the analysis. 4. Labelling The process of assigning a category or class label to each record based on specific criteria. In this study the text labels used were positive, negative, or neutral, done manually or using pre-trained algorithms. 5. TF-IDF Term Frequency-Inverse Document Frequency (TF-IDF) is employed to transform text into a vector format. This statistical measure evaluates the importance of a word within a corpus, providing a weighting factor that enhances the effectiveness of the Naïve Bayes model in distinguishing relevant terms in tweets. 6. Model Training a. Naive Bayes Model Training The process of training the Naive Bayes classification algorithm with labeled datasets. The model uses probability and statistics to predict sentiment labels from text, trained with datasets that have been supplemented with sentiment labels in the previous step b. GBM Model Training The process trains a Gradient Boosted Machines model, which is an advanced classification algorithm that uses boosting techniques to generate predictive models. Explanation: GBM builds a series of weak predictive models incrementally, with each new model attempting to correct errors from the previous model. 7. Model Evaluation The process of testing a trained model against a test dataset to assess model performance using metrics such as accuracy, precision, recall, and F1 scores through "performance" and "cross validation" methods. Explanation: This metric gives an idea of how well the model can classify new data and how far the predicted results correspond to actual values. 8. Analysis of Results This process performs an examination of the classification results, looking for relationships, relationships and patterns in the data predicted from the model created 9. Analysis Tools This study used RapidMiner as the main tool for data analysis. RapidMiner is an advanced and user-friendly data analysis platform, which allows users to perform various analytical processes without the need for in-depth programming knowledge [24]. This includes data preprocessing, modeling, and model evaluation. IV. RESULTS The pre-processing phase of the data begins with reading a CSV file containing 9,952 entries that are comments from Twitter. This step is important to ensure that all necessary data is accessible and ready for the next analysis process. After that, it continues with the sentiment labeling stage where each comment is categorized into one of three sentiment labels: positive, negative, or neutral. This process is essential for opinion mining or further sentiment analysis, allowing for a classification based on actual data and ongoing situations. The next process carried out is the conversion process from nominal data to text to facilitate tokenization. Tokenization successfully breaks down text into smaller units, or tokens, allowing for a more detailed analysis of the individual components of that text. A case transformation is then performed, converting all text to lowercase form to avoid unnecessary duplication of tokens based on capitalization differences. The next step is stopword removal using a collection of stopword Indonesian downloaded from the Kaggle site, which eliminates words that do not carry significant sentiment weight and helps simplify the dataset. Lastly, token filters are applied by length to weed out tokens that are too short or too long that may not be relevant for sentiment analysis. This pre-processing process produces a clean, structured dataset ready for the model training phase using Naive Bayes and Gradient Boosted Machines. 1. Model Training Model training includes training on Naive Bayes models and Gradient Boosted Machines. The dataset used for Naive Bayes model training has been processed using the TF-IDF (Term Frequency-Inverse Document Frequency) method, which is commonly used to convert text into a format that can be understood by machine learning models. The dataset consists of 8,477 examples with 11,510 regular attributes. 2. Evaluation Analysis Evaluate the model using the cross validation method. This is a very common and robust validation method, where data is divided into "folds" or parts. In each iteration, one fold is used as test data and the rest as training data. The following are the cross validation results for each model 3. Naïve Bayes: Table 1 Model Evaluation Naive Bayes True positive True negative True neutral Class Precision Pred. positive 1310 13 12 98.13% Pred. negative 19 32 12 50.79% Pred. neutral 12 15 39 59.09% Class recall 97.69% 53.33% 61.90% The results of cross validation for the Naïve Bayes model showed a fairly good classification performance in the 'positive' class with 98.13% precision and 97.69% recall , indicating that the model is very effective in identifying and predicting positive sentiment with a very low error rate. For the 'negative' and 'neutral' classes, precision and recall showed lower performance. The precision for 'negative' was 50.79% and for 'neutral' 59.09%, while the recall for those classes was 53.33% and 61.90% respectively. The lower precision in these two classes suggests that the model has difficulty in consistently identifying negative and neutral comments, often classifying those comments incorrectly. A moderate recall indicates that while the model could identify most negative and neutral comments, there was still a significant amount missed. 4. Gradient Boosted Machines Table 2 GBM Model Evaluation tTrue positive tTrue negative trTue neutral CClass Precision Ppred. positive 1298 29 11 97.01% Ppred. negative 30 23 4 40.35% Ppred. neutral 13 8 48 69.57% CClass recall 96.79% 38.33% 76.19% The results of cross validation for the Gradient Boosted Machines (GBM) model showed excellent classification performance in the 'positive' class with 97.01% precision and 96.79% recall. This performance indicates that the model is very effective in identifying and predicting positive sentiment with minimal error rates. However, for the 'negative' and 'neutral' classes, the results showed lower levels of effectiveness. The precision for the 'negative' classes was 40.35% and for the 'neutral' was 69.57%, while the recall for those classes was 38.33% and 76.19%, respectively. The relatively low precision for the 'negative' class indicates that models are often wrong in identifying negative comments, tending to classify comments that are actually positive or neutral as negative. This was also followed by low recall, suggesting that the model failed to capture a large number of actual negative comments. In the 'neutral' class, although relatively higher recall indicates that the model is better at capturing neutral comments, a precision lower than 70% still indicates an error in classifying comments from other categories as neutral. Overall, while the GBM model shows excellent capabilities in managing positive categories, there is significant room for improvement in the handling of negative and neutral categories, particularly in improving precision and reducing classification errors. 5. F1 Score To add the F1 score to the comparison table for the Naive Bayes and Gradient Boosted Machines (GBM) model, it can be calculated through the following formula: The results of the F1 score calculation from both models are as follows: Table 3 F1 Score from Prediction Model Sentiment Categories Metric Naive Bayes Gradient Boosted Machines Positive F1 Score 97.91% 96.90% Negative F1 Score 52.03% 39.31% Neutral F1 Score 60.47% 72.71% V. DISCUSSION 1. Model Performance Evaluation The performance of the Naive Bayes and Gradient Boosted Machines (GBM) models in sentiment classification reveals their respective strengths and limitations. Consistent with findings from previous research [21], the Naive Bayes model excelled in identifying positive sentiments with high precision and recall, indicative of its robust pattern recognition capabilities for positively connotated words [10]. However, it showed limitations in classifying negative and neutral sentiments, echoing challenges noted in other sentiment analyses using Naive Bayes [11], [12]. This model’s difficulty with these categories could be due to its probabilistic nature, which may not capture the contextual nuances essential in these more complex sentiment classifications [22]. Conversely, the GBM model displayed a commendable balance across all sentiment categories, particularly excelling in identifying neutral sentiments [23]. This capability suggests that GBM is adept at handling the ambiguity and subtlety often associated with neutral expressions, potentially due to its iterative refinement process that enhances accuracy over successive models [15], [16]. Despite a slight decline in precision for positive sentiments compared to Naive Bayes, GBM’s performance remained robust, showcasing its utility in applications requiring nuanced sentiment detection [14]. 2. Methodological Analysis The application of TF-IDF for data preprocessing has been validated as an effective method for preparing text for analysis [13], [17]. However, the challenge of managing a large number of attributes (11,510), as noted in our findings, is consistent with other studies indicating potential overfitting issues in Naive Bayes when faced with high-dimensional data [18]. Future research might explore dimensionality reduction strategies, such as those discussed by Dang et al. [15], to enhance model performance without losing critical information. 3. Influence and Implications of Research Results Our research contributes valuable insights into public sentiments regarding judicial decisions, aligning with the broader discourse on political communications and public opinion [2], [3], [4] The ability of the models, particularly Naive Bayes, to effectively classify positive sentiments aligns with their potential utility in monitoring public reactions to governmental decisions, a critical aspect of democratic governance [6], [9]. However, the noted deficiencies in identifying negative and neutral sentiments highlight the ongoing need for methodological advancements to refine sentiment analysis techniques, ensuring a comprehensive understanding of public opinion [19], [20]. This study underscores the importance of continuous improvement in sentiment analysis methodologies to achieve balanced and accurate insights, which are essential for policymakers, social scientists, and political analysts aiming to gauge public opinion and adapt strategies accordingly [24]. VI. CONCLUSION This study has effectively demonstrated the use of Naive Bayes and Gradient Boosted Machines (GBM) models in performing sentiment analysis of Twitter data related to the Constitutional Court’s decision on the 2024 Indonesian elections. Our findings underscore the Naive Bayes model’s strong performance in accurately identifying positive sentiments, affirming its utility in applications where positive sentiment detection is crucial. However, the model exhibited limitations in accurately classifying negative and neutral sentiments, which underscores the necessity for enhancements in handling complex sentiment classifications. The GBM model, on the other hand, showed a more balanced performance across all sentiment categories. It was particularly adept at recognizing neutral sentiments, highlighting its capacity to manage the subtleties and ambiguities of such expressions. This balanced performance advocates for the use of GBM in scenarios that require nuanced sentiment detection across a broader spectrum. Methodologically, the application of the TF-IDF technique proved to be effective for text preprocessing, though the challenge of handling a large number of attributes suggests that further improvements could be made through dimensionality reduction techniques. Such adjustments could potentially enhance the generalizability and efficiency of the models without compromising the depth of analysis. The implications of our study are significant for stakeholders interested in public opinion analysis, including policymakers, political analysts, and social media strategists. By improving sentiment analysis techniques, a more comprehensive and accurate understanding of public sentiments can be achieved, which is pivotal for informed decision-making. Future work should focus on refining these models to enhance their ability to classify negative and neutral sentiments more accurately. Additionally, exploring other machine learning techniques and incorporating more sophisticated natural language processing tools could further improve the robustness and accuracy of sentiment analysis in political and social contexts. Declarations The authors declare that they have no competing interests, either financial or non-financial, that could have appeared to influence the work reported in this paper. FUNDING This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. AKNOWLEDGEMENT We would like to express our gratitude to everyone who contributed to the successful completion of this research. Special thanks go to the faculty and staff of the Department of Computer Science at Institut Mahardika, whose insights and expertise have been invaluable throughout this project. We are particularly grateful for the guidance and support provided by Dr. Christina Juliane, whose expertise in the field of data science significantly enhanced our analytical approaches. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. However, the resources and facilities provided by Institut Mahardika have been instrumental in carrying out this study. Author Contribution Indra Surya Permana and Fardhoni led the main research activities, including conceptualization, methodology design, data collection, analysis, and initial manuscript drafting. Indra Surya Permana, serving as the corresponding author, additionally handled the project administration, supervision, and prepared the final version of the manuscript, integrating contributions and feedback. Fardhoni contributed specifically to data curation and provided substantial revisions to the manuscript drafts. Christina Juliane supervised the overall project, providing critical academic guidance and intellectual input, and reviewed the manuscript to ensure the accuracy and integrity of the work. All authors read and approved the final manuscript, agreeing to be accountable for all aspects of the work. References Afnira E, Optimalisasi Media Sosial sebagai Sarana Publikasi Pengawasan Pemilu (2024) : Kasus Bawaslu Kota Tanjungpinang, J. Mhs. Komun. Cantrik , vol. 3, no. 1, Art. no. 1, May 2023, 10.20885/cantrik.vol3.iss1.art4 Indonesia BBCN, Putusan MK Reaksi setelah MK tolak gugatan hasil Pilpres 2024, dari Prabowo hingga pengunjuk rasa, BBC News Indonesia. Accessed: Apr. 23, 2024. [Online]. Available: https://www.bbc.com/indonesia/articles/ce437lrlzwlo Hutabarat J Menjinakkan Polarisasi dan Politik Identitas Jelang Pemilu 2024, tirto.id. Accessed: Apr. 23, 2024. [Online]. Available: https://tirto.id/menjinakkan-polarisasi-dan-politik-identitas-jelang-pemilu-2024-gtQ8 Nur MF Potensi Polarisasi & Mencegah Konflik Pendukung di Pemilu 2024, tirto.id. Accessed: Apr. 23, 2024. [Online]. Available: https://tirto.id/potensi-polarisasi-mencegah-konflik-pendukung-di-pemilu-2024-gRPS Indonesia KPU Daftar Sengketa Pemilu Presiden & Wakil Presiden Ri 2024: Daftar Daerah Yang Mengalami Pemungutan Dan Penghitungan Suara Ulang/Susulan/Lanjutan (Baik Seluruhnya Maupun Sebagian), Komisi Pemilihan Umum. Accessed: Apr. 23, 2024. [Online]. Available: https://pemilu2024.kpu.go.id/pilpres/daftar-sengketa Dhiman A, Toshniwal D (Oct. 2022) AI-based Twitter framework for assessing the involvement of government schemes in electoral campaigns. Expert Syst Appl 203:117338. 10.1016/j.eswa.2022.117338 Devi WR, Chingangbam C (2021) Sentiment Analysis for Electoral Prediction Using Twitter Data, in Emerging Technologies in Data Mining and Information Security , vol. 1300, A. E. Hassanien, S. Bhattacharyya, S. Chakrabati, A. Bhattacharya, and S. Dutta, Eds., in Advances in Intelligent Systems and Computing, vol. 1300., Singapore: Springer Nature Singapore, pp. 25–33. 10.1007/978-981-33-4367-2_3 Vindua R, Zailani AU (2023) Analisis Sentimen Pemilu Indonesia Tahun 2024 Dari Media Sosial Twitter Menggunakan Python, Jur. Ris. Kom. , vol. 10, no. 2, p. 479, Apr. 10.30865/jurikom.v10i2.5945 Septian L, Aljauza T, Juliane C (2023) Analisis Sentimen Putusan Mahkamah Konstitusi terhadap Batas Usia Capres dan Cawapres Menggunakan IndoBERT, Indonesian Journal of Computer Science , vol. 12, no. 6, Art. no. 6, 10.33022/ijcs.v12i6.3614 Isnain AR, Marga NS, Alita D, Sentiment Analysis Of Government Policy On Corona Case Using Naive Bayes Algorithm, (Jan. 2021) Indonesian J Comput Cybern Syst 15(1):55. 10.22146/ijccs.60718 Zahra DF (2023) Analisis Sentimen penggunaan Mypertamina untuk Pembelian BBM Bersubsidi mengggunakan Algoritma Naive Bayes, Explore. jurnal. sistem. inf. dan. telematika , vol. 14, no. 1, p. 94, Jul. 10.36448/jsit.v14i1.3098 Ansari S, Hans K, Khatri SK (2017) A Naive Bayes classifier approach for detecting hypervisor attacks in virtual machines, 2nd International Conference on Telecommunication and Networks, TEL-NET vol. 2018-January, pp. 1–6, Apr. 2018, 10.1109/TEL-NET.2017.8343551 Duan HK, Vasarhelyi MA, Codesso M, Alzamil Z (2023) Enhancing the government accounting information systems using social media information: An application of text mining and machine learning, International Journal of Accounting Information Systems , vol. 48, p. 100600, Mar. 10.1016/j.accinf.2022.100600 Nair A, Paralkar C, Pandya J, Chopra Y, Krishnan D (2021) Comparative Review on Sentiment analysis-based Recommendation system, in 6th International Conference for Convergence in Technology (I2CT) , Apr. 2021, pp. 1–6. 10.1109/I2CT51068.2021.9418222 Dang NC, Moreno-García MN, De la Prieta F (2020) Sentiment Analysis Based on Deep Learning: A Comparative Study, Electronics , vol. 9, no. 3, Art. no. 3, Mar. 10.3390/electronics9030483 Mehmood K, Essam D, Shafi K, Malik MK (2019) Sentiment Analysis for a Resource Poor Language—Roman Urdu, ACM Trans. Asian Low-Resour. Lang. Inf. Process. , vol. 19, no. 1, p. 10:1–10:15, Aug. 10.1145/3329709 Gogia E (Dec. 2021) A Review on Sentiment Analysis techniques and applications. GRET 24–29. 10.47893/GRET.2021.1041 Kuckartz U (2014) Qualitative text analysis: a guide to methods, practice & using software. Sage, Los Angeles London New Delhi Singapore Washington, DC Mohan Kumar AV, Nandkumar AN (May 2020) A Survey on Challenges and Research Opportunities in Opinion Mining. SN COMPUT SCI 1(3):171. 10.1007/s42979-020-00149-4 Osmani A, Mohasefi JB (2023) Opinion Mining Using Enriched Joint Sentiment-Topic Model, Int. J. Info. Tech. Dec. Mak. , vol. 22, no. 01, pp. 313–375, Jan. 10.1142/S0219622022500584 Foster J, Wagner J (2021) Naive Bayes versus BERT: Jupyter notebook assignments for an introductory NLP course, in Proceedings of the Fifth Workshop on Teaching NLP , Online: Association for Computational Linguistics, pp. 112–114. 10.18653/v1/2021.teachingnlp-1.20 Alzamzami F, Hoda M, Saddik AE (2020) Light Gradient Boosting Machine for General Sentiment Classification on Short Texts: A Comparative Evaluation. IEEE Access 8:101840–101858. 10.1109/ACCESS.2020.2997330 Khalid M, Ashraf I, Mehmood A, Ullah S, Ahmad M, Choi GS (2020) GBSVM: Sentiment Classification from Unstructured Reviews Using Ensemble Classifier, Applied Sciences , vol. 10, no. 8, p. 2788, Apr. 10.3390/app10082788 Makhi A, Fauzan AC, SISTEM INFORMASI MANAJEMEN ADMINISTRASI PADA, Apr. LEMBAGA SERTIFIKASI PROFESI P1 UNIVERSITAS NAHDLATUL ULAMA BLITAR MENGGUNAKAN MODEL RAPID APPLICATION DEVELOPMENT (2022), jatim , vol. 3, no. 1, pp. 21–34, 10.31102/jatim.v3i1.1423 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4482093","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":309622328,"identity":"ca165287-09fa-4001-a169-149f2644214e","order_by":0,"name":"Indra Surya Permana","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYBACNgYGAwYGHiCLvQFIGFiQooXnAEiLBFEWGUAoiQQwSVg9n3Tztg8/ZOyi5Wc+v7rhR4EEA397dwJ+h8kcK57Zw5Oc2zg7p+xmD9BhEmfObsCvRSLHmIGHhzm3WTon7QYPUIuBRC5hLYx/eOpz2yTPpN38Q6wWZh6ew7k9EuzHbhNpS1oxswzP8dwZPDlst2UMJHgI+kV+RvJmxrc91bnz248/u/nmj40cf3svfi1gwNgDInnAEcRDWDkY/AAR7A+IVD0KRsEoGAUjDQAABcM/6CwqyWgAAAAASUVORK5CYII=","orcid":"","institution":"Institut Teknologi dan Kesehatan Mahardika","correspondingAuthor":true,"prefix":"","firstName":"Indra","middleName":"Surya","lastName":"Permana","suffix":""},{"id":309622329,"identity":"ac79e79f-b58e-4469-8b55-d7eb5176ee17","order_by":1,"name":"Fardhoni Fardhoni","email":"","orcid":"","institution":"Institut Teknologi dan Kesehatan Mahardika","correspondingAuthor":false,"prefix":"","firstName":"Fardhoni","middleName":"","lastName":"Fardhoni","suffix":""},{"id":309622330,"identity":"1182df64-f5d7-456b-ac56-d895b1d67f92","order_by":2,"name":"Christina Juliane","email":"","orcid":"","institution":"STMIK LIKMI School of Business \u0026 IT","correspondingAuthor":false,"prefix":"","firstName":"Christina","middleName":"","lastName":"Juliane","suffix":""}],"badges":[],"createdAt":"2024-05-27 03:53:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4482093/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4482093/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":57887194,"identity":"74ecf52c-299f-4e42-8358-606f49387654","added_by":"auto","created_at":"2024-06-07 05:17:39","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":52387,"visible":true,"origin":"","legend":"\u003cp\u003eResearch Flowchart\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4482093/v1/cca82340bad6f51103c5eedd.png"},{"id":57887857,"identity":"23ad55b8-2857-45cb-bb0b-d051139762df","added_by":"auto","created_at":"2024-06-07 05:33:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":421868,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4482093/v1/926f2fd9-c2e9-4630-a108-139f2ddbac7b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003ePublic Response to the Constitutional Court’s Decision on Indonesia’s 2024 Elections\u003c/p\u003e","fulltext":[{"header":"I. INTRODUCTION","content":"\u003cp\u003eThe 2024 General Election in Indonesia is an important event that attracts great attention from the public and is in the spotlight. In this context, several important dynamics have already occurred. First, technology, especially social media, has become the main platform for presidential candidates and their supporters to convey their vision, mission, and work programs [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Second, political education and understanding the impact of political decisions are becoming increasingly important in the context of voter participation. Third, global issues such as climate change, energy security, and global health are the main focus in candidate debates [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Fourth, the election, especially the 2024 Presidential Election, reflects the spirit of inclusivity with many candidates from various backgrounds and community groups. Fifth, political polarization, potential abuse of authority, and the threat of fertile dynastic politics are issues that color the course of the 2024 presidential election. Sixth, young people aged 22\u0026ndash;30 years will dominate the electorate nationwide, with a share of 56%, or about 114\u0026nbsp;million. Seventh, misinformation and disinformation on social media are still major issues ahead of the 2024 elections [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe 2024 election has been completed when the results of the recapitulation are announced in March 2024, where the presidential candidate and vice president pair number 2 get 96,214,691 votes or more than 57% of the total votes and win a victory that is far adrift from the pair number 1 and 2. however, the results of the 2024 election in Indonesia have caused various controversies and debates, including the existence of 718 disputes recorded in the 2024 Presidential \u0026amp; Vice Presidential Election Dispute List [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. The Constitutional Court (MK) decided to reject all claims over the results of the 2024 Presidential Election. However, many parties felt disappointed and doubted the decision.\u003c/p\u003e \u003cp\u003eCurrently, platform X or Twitter is often considered to be a source of information in the analysis of public opinion [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. This is because Twitter is a place for people to share opinions, express emotional conditions such as happiness, disappointment, anger, and so on, including political views of a person or group [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe opinions of social media user X (Twitter) regarding the results of the Constitutional Court (MK) decision on the 2024 election are very diverse and reflect the existing political dynamics. Some Twitter users showed their support for the Constitutional Court's decision, while others expressed disappointment and doubt. There were also a number of users who questioned the process and legal basis of the decision. For example, some have questioned the independence of the election administration agency and highlighted allegations of abuse of power by the government. The behavior of social media users is discussed in research conducted by Septian, et al [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] and Vindua and Zailani [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] which discusses the sentiment of Twitter users towards a legal ruling in Indonesia on a national scale.\u003c/p\u003e \u003cp\u003eIn the digital age, the abundance of sentiment data from social media and online platforms provides a valuable source of information to analyze public views to see the sentiments or opinions formed. However, the complexity and volume of this data demands a sophisticated approach to its analysis. This is where Naive Bayes models and Gradient Boosted Machines play an important role in opinion mining. Naive Bayes, with its prominence in text classification, has become one of the popular approaches in sentiment analysis as has previous research conducted by Isnain, et al [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], Zahra [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], and Ansari [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. While Gradient Boosted Machines (GBM) models are powerful methods in machine learning that can produce accurate prediction models [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. In the context of sentiment analysis, gradient boosted machines can be used to model non-linear relationships between features in text and existing sentiment[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eUsing these models, opinion mining or sentiment analysis of the Constitutional Court's decision related to the 2024 election can be carried out to identify patterns and trends in public opinion. The sentiment analysis can provide an idea of how the public responded to the Constitutional Court's decision, whether the majority supported or questioned the decision, and what factors influenced their views.\u003c/p\u003e \u003cp\u003e \u003cb\u003eResearch Objectives\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo find out User X's (Twitter) Opinion Mining on the Constitutional Court Ruling on the 2024 Election with Naive Bayes and Gradient Boosted Machines models.\u003c/p\u003e"},{"header":"II. LITERATURE REVIEW","content":"\u003cp\u003eText Analysis in Natural Language Processing (NLP) is one of the methods in machine learning (machine learning) used to understand and extract information from text [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. This method involves processing and reading text using computer algorithms to recognize patterns and look up the meaning of the text. These methods of text analysis can include identification of entities such as names of people or places, classification of text into specific categories, information extraction, sentiment analysis, and others.\u003c/p\u003e \u003cp\u003eOpion mining or often called Sentiment analysis is a method used in text analysis to identify, extract, and analyze sentiment or opinion. Opion mining is important because with the increasing complexity and volume of data generated by social media and online platforms, it is necessary to understand public opinion quickly and accurately with the help of machine learning [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSome methods in machine learning go into simple methods and also deep learning or deep learning. One method in opinion mining that is often done is to use the Na\u0026iuml;ve Bayes Classifier (NBC) method. The NBC method itself is a simple and efficient method that can be used to classify text into different sentiment categories [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. In the Naive Bayes method, probabilistics are used to classify text into sentiment categories based on the likelihood of a sentiment occurring based on the linguistic features present in the text.\u003c/p\u003e \u003cp\u003eGradient Boosted Machines (GBM) are machine learning techniques that use a weak set of predictive models (usually decision trees) to build powerful predictive models incrementally [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. The algorithm improves prediction by iteratively adjusting a new model that focuses on cases incorrectly predicted by previous models in the series. Gradient Boosted Machines (GBM) models are sophisticated machine learning techniques for regression and classification. It belongs to the category of 'ensemble learning' algorithms, where multiple predictive models are combined to improve the accuracy and power of predictions [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBoth NBC and GBM have distinct roles in sentiment analysis. While Na\u0026iuml;ve Bayes offers speed and simplicity, making it suitable for initial screenings of sentiment in large datasets, GBM provides depth and precision, particularly valuable in contexts where the differentiation of subtle sentiment nuances is critical. Together, these models contribute to a more holistic sentiment analysis process, allowing for both broad and deep analyses of public opinions on platforms like Twitter regarding critical events such as the Constitutional Court decisions i.e regarding Indonesian 2024 electoral decision [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe relevance of these models extends to analyzing public reactions to significant societal events, such as judicial decisions. By applying these sentiment analysis models, researchers can capture a broad spectrum of public opinions, providing valuable insights into societal perceptions and influencing future policy formulations [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. This is particularly pertinent in the context of electoral decisions or constitutional judgments, where public sentiment can significantly impact the sociopolitical landscape.\u003c/p\u003e \u003cp\u003eThe following \u0026ldquo;State of the Art\u0026rdquo; table synthesizes significant studies in this area, summarizing their methodologies, key findings, and contributions to the discipline. The table aims to highlight how various approaches have been utilized to enhance our understanding of public opinion and sentiment, reflecting the diverse applications of these technologies across different research contexts\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eState of The Art\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStudy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMethodology\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKey Findings\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eContribution\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e], [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial media monitoring, AI-based Twitter analysis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eShowed social media's role in election oversight and developed AI frameworks for evaluating government campaign strategies.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eHighlighted social media and AI's roles in political monitoring and strategy assessment.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDescriptive analysis, Socio-political studies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDescribed public reactions to judicial decisions and examined electoral conflicts and polarization.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOffered insights into public responses and political dynamics surrounding elections.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSentiment analysis with Twitter data, IndoBERT application\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eApplied sentiment analysis to predict election results and analyze legal sentiments.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eShowcased sentiment analysis's utility in predicting political outcomes and legal reactions.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNa\u0026iuml;ve Bayes for sentiment analysis on government policies and education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEvaluated public sentiments on policies and educational initiatives, demonstrating broad effectiveness.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEnhanced understanding of public responses to governmental and educational sectors through NLP.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGBM and ensemble learning\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDiscussed GBM's role in complex classifications and its use to boost prediction accuracy.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDemonstrated GBM's capability in detailed sentiment analysis and improved prediction reliability.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMachine learning in opinion mining, enriched joint sentiment-topic models\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInvestigated opinion mining challenges and introduced new joint sentiment-topic models.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIntroduced innovative methods for detailed opinion analysis from vast datasets.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"},{"header":"III. METHODS","content":"\u003cp\u003eThis research employs opinion mining techniques using the Naïve Bayes Classifier (NBC) and Gradient Boosted Machines (GBM) to analyze sentiments expressed on Twitter regarding the Constitutional Court’s decision on the 2024 Indonesian Presidential Election. The NBC method is a probabilistic classifier known for its simplicity and effectiveness in text classification, which categorizes text based on the likelihood of sentiment occurrence, considering the linguistic features present in the text. Concurrently, GBM enhances prediction accuracy by iteratively refining models that focus on misclassified cases by previous iterations.\u003c/p\u003e\n\u003cp\u003eThe following are the stages that will be carried out in sentiment analysis in this study:\u003c/p\u003e\n\u003cp\u003e1.\u0026nbsp; \u0026nbsp;Data Collection\u003c/p\u003e\n\u003cp\u003eThe process of collecting data relevant to the research. In this context, it involves collecting tweets or text from platform X related to the research topic.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.\u0026nbsp; \u0026nbsp;Dataset\u003c/p\u003e\n\u003cp\u003eThe dataset used is a collection of Platform X (or Twitter) comments that use hashtags related to the Constitutional Court's decision trending 5 days before the day when the constitutional court's final decision regarding the Dispute over the Results of the 2024 Presidential and Vice Presidential Elections on April 22, 2024. The dataset is all trending topics with queries BarengPrabowoBangunBangsa, Gibran, Constitutional Court, Constitutional Court Verdict, PeopleWin, Anies Muhaimin, and Ganjar Mahfud. The total data that can be retrieved is 11,721 lines that include username, tweet ID, date, Tweet text content, and source.\u003c/p\u003e\n\u003cp\u003e3.\u0026nbsp; \u0026nbsp;Data Pre-processing\u003c/p\u003e\n\u003cp\u003eThe stage of preparing the raw text for analysis by cleaning it of unnecessary elements or not providing important information. These include:\u003c/p\u003e\n\u003cp\u003e1.\u0026nbsp; \u0026nbsp;Case Folding: Normalizing all text to lowercase to ensure uniformity.\u003c/p\u003e\n\u003cp\u003e2.\u0026nbsp; \u0026nbsp;Filtering: Removing non-essential characters like numerals and symbols to focus purely on textual data.\u003c/p\u003e\n\u003cp\u003e3.\u0026nbsp; \u0026nbsp;Stopword Removal: Eliminating common words that offer minimal unique information for analysis.\u003c/p\u003e\n\u003cp\u003e4.\u0026nbsp; \u0026nbsp;Tokenizing: Segmenting text into individual words or tokens.\u003c/p\u003e\n\u003cp\u003e5.\u0026nbsp; \u0026nbsp;Stemming: Reducing words to their base or root form to simplify the analysis.\u003c/p\u003e\n\u003cp\u003e4.\u0026nbsp; \u0026nbsp;Labelling\u003c/p\u003e\n\u003cp\u003eThe process of assigning a category or class label to each record based on specific criteria. In this study the text labels used were positive, negative, or neutral, done manually or using pre-trained algorithms.\u003c/p\u003e\n\u003cp\u003e5.\u0026nbsp; \u0026nbsp;TF-IDF\u003c/p\u003e\n\u003cp\u003eTerm Frequency-Inverse Document Frequency (TF-IDF) is employed to transform text into a vector format. This statistical measure evaluates the importance of a word within a corpus, providing a weighting factor that enhances the effectiveness of the Naïve Bayes model in distinguishing relevant terms in tweets.\u003c/p\u003e\n\u003cp\u003e6.\u0026nbsp; \u0026nbsp;Model Training\u003c/p\u003e\n\u003cp\u003ea.\u0026nbsp; \u0026nbsp;\u0026nbsp;Naive Bayes Model Training\u003c/p\u003e\n\u003cp\u003eThe process of training the Naive Bayes classification algorithm with labeled datasets. The model uses probability and statistics to predict sentiment labels from text, trained with datasets that have been supplemented with sentiment labels in the previous step\u003c/p\u003e\n\u003cp\u003eb.\u0026nbsp; \u0026nbsp;\u0026nbsp;GBM Model Training\u003c/p\u003e\n\u003cp\u003eThe process trains a Gradient Boosted Machines model, which is an advanced classification algorithm that uses boosting techniques to generate predictive models. Explanation: GBM builds a series of weak predictive models incrementally, with each new model attempting to correct errors from the previous model.\u003c/p\u003e\n\u003cp\u003e7.\u0026nbsp; \u0026nbsp;Model Evaluation\u003c/p\u003e\n\u003cp\u003eThe process of testing a trained model against a test dataset to assess model performance using metrics such as accuracy, precision, recall, and F1 scores through \"performance\" and \"cross validation\" methods. Explanation: This metric gives an idea of how well the model can classify new data and how far the predicted results correspond to actual values.\u003c/p\u003e\n\u003cp\u003e8.\u0026nbsp; \u0026nbsp;Analysis of Results\u003c/p\u003e\n\u003cp\u003eThis process performs an examination of the classification results, looking for relationships, relationships and patterns in the data predicted from the model created\u003c/p\u003e\n\u003cp\u003e9.\u0026nbsp; \u0026nbsp;Analysis Tools\u003c/p\u003e\n\u003cp\u003eThis study used RapidMiner as the main tool for data analysis. RapidMiner is an advanced and user-friendly data analysis platform, which allows users to perform various analytical processes without the need for in-depth programming knowledge\u0026nbsp;[24]. This includes data preprocessing, modeling, and model evaluation.\u0026nbsp;\u003c/p\u003e"},{"header":"IV. RESULTS","content":"\u003cp\u003eThe pre-processing phase of the data begins with reading a CSV file containing 9,952 entries that are comments from Twitter. This step is important to ensure that all necessary data is accessible and ready for the next analysis process. After that, it continues with the sentiment labeling stage where each comment is categorized into one of three sentiment labels: positive, negative, or neutral. This process is essential for opinion mining or further sentiment analysis, allowing for a classification based on actual data and ongoing situations.\u003c/p\u003e\n\u003cp\u003eThe next process carried out is the conversion process from nominal data to text to facilitate tokenization. Tokenization successfully breaks down text into smaller units, or tokens, allowing for a more detailed analysis of the individual components of that text. A case transformation is then performed, converting all text to lowercase form to avoid unnecessary duplication of tokens based on capitalization differences.\u003c/p\u003e\n\u003cp\u003eThe next step is stopword removal using a collection of stopword Indonesian downloaded from the Kaggle site, which eliminates words that do not carry significant sentiment weight and helps simplify the dataset. Lastly, token filters are applied by length to weed out tokens that are too short or too long that may not be relevant for sentiment analysis. This pre-processing process produces a clean, structured dataset ready for the model training phase using Naive Bayes and Gradient Boosted Machines.\u003c/p\u003e\n\u003cp\u003e1. \u0026nbsp; Model Training\u003c/p\u003e\n\u003cp\u003eModel training includes training on Naive Bayes models and Gradient Boosted Machines. The dataset used for Naive Bayes model training has been processed using the TF-IDF (Term Frequency-Inverse Document Frequency) method, which is commonly used to convert text into a format that can be understood by machine learning models. The dataset consists of 8,477 examples with 11,510 regular attributes.\u003c/p\u003e\n\u003cp\u003e2. \u0026nbsp; Evaluation\u003c/p\u003e\n\u003cp\u003eAnalysis Evaluate the model using the cross validation method. This is a very common and robust validation method, where data is divided into \u0026quot;folds\u0026quot; or parts. In each iteration, one fold is used as test data and the rest as training data. The following are the cross validation results for each model\u003c/p\u003e\n\u003cp\u003e3. \u0026nbsp; Na\u0026iuml;ve Bayes:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTable\u0026nbsp;\u003c/em\u003e\u003cem\u003e1\u003c/em\u003e\u003cem\u003e\u0026nbsp;Model Evaluation Naive Bayes\u003c/em\u003e\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTrue positive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTrue negative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTrue neutral\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClass Precision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePred. positive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e98.13%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePred. negative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e50.79%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePred. neutral\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e59.09%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClass recall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e97.69%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e53.33%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e61.90%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u0026nbsp;The results of cross validation for the Na\u0026iuml;ve Bayes model showed a fairly good classification performance in the \u0026apos;positive\u0026apos; class with \u003cstrong\u003e98.13% precision\u003c/strong\u003e and \u003cstrong\u003e97.69% recall\u003c/strong\u003e, indicating that the model is very effective in identifying and predicting positive sentiment with a very low error rate. For the \u0026apos;negative\u0026apos; and \u0026apos;neutral\u0026apos; classes, precision and recall showed lower performance. The precision for \u0026apos;negative\u0026apos; was \u003cstrong\u003e50.79%\u003c/strong\u003e and for \u0026apos;neutral\u0026apos; \u003cstrong\u003e59.09%,\u003c/strong\u003e while the recall for those classes was \u003cstrong\u003e53.33%\u003c/strong\u003e and \u003cstrong\u003e61.90%\u003c/strong\u003e respectively. The lower precision in these two classes suggests that the model has difficulty in consistently identifying negative and neutral comments, often classifying those comments incorrectly. A moderate recall indicates that while the model could identify most negative and neutral comments, there was still a significant amount missed.\u003c/p\u003e\n\u003cp\u003e4. \u0026nbsp; Gradient Boosted Machines\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTable\u0026nbsp;\u003c/em\u003e\u003cem\u003e2\u003c/em\u003e\u003cem\u003e\u0026nbsp;GBM Model Evaluation\u003c/em\u003e\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003etTrue positive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003etTrue negative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003etrTue neutral\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCClass Precision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePpred. positive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1298\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e97.01%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePpred. negative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e40.35%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePpred. neutral\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e69.57%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCClass recall\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e96.79%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e38.33%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e76.19%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u0026nbsp;The results of cross validation for the Gradient Boosted Machines (GBM) model showed excellent classification performance in the \u0026apos;positive\u0026apos; class with 97.01% precision and 96.79% recall. This performance indicates that the model is very effective in identifying and predicting positive sentiment with minimal error rates. However, for the \u0026apos;negative\u0026apos; and \u0026apos;neutral\u0026apos; classes, the results showed lower levels of effectiveness. The precision for the \u0026apos;negative\u0026apos; classes was 40.35% and for the \u0026apos;neutral\u0026apos; was 69.57%, while the recall for those classes was 38.33% and 76.19%, respectively.\u003c/p\u003e\n\u003cp\u003eThe relatively low precision for the \u0026apos;negative\u0026apos; class indicates that models are often wrong in identifying negative comments, tending to classify comments that are actually positive or neutral as negative. This was also followed by low recall, suggesting that the model failed to capture a large number of actual negative comments. In the \u0026apos;neutral\u0026apos; class, although relatively higher recall indicates that the model is better at capturing neutral comments, a precision lower than 70% still indicates an error in classifying comments from other categories as neutral.\u003c/p\u003e\n\u003cp\u003eOverall, while the GBM model shows excellent capabilities in managing positive categories, there is significant room for improvement in the handling of negative and neutral categories, particularly in improving precision and reducing classification errors.\u003c/p\u003e\n\u003cp\u003e5. \u0026nbsp; F1 Score\u003c/p\u003e\n\u003cp\u003eTo add the F1 score to the comparison table for the Naive Bayes and Gradient Boosted Machines (GBM) model, it can be calculated through the following formula:\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eThe results of the F1 score calculation from both models are as follows:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eTable\u0026nbsp;\u003c/em\u003e\u003cem\u003e3\u003c/em\u003e\u003cem\u003e\u0026nbsp;F1 Score from Prediction Model\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"27.551020408163264%\" valign=\"top\"\u003e\n \u003cp\u003eSentiment Categories\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eMetric\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.387755102040817%\" valign=\"top\"\u003e\n \u003cp\u003eNaive Bayes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"38.775510204081634%\" valign=\"top\"\u003e\n \u003cp\u003eGradient Boosted Machines\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"27.551020408163264%\" valign=\"top\"\u003e\n \u003cp\u003ePositive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eF1 Score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.387755102040817%\" valign=\"top\"\u003e\n \u003cp\u003e97.91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"38.775510204081634%\" valign=\"top\"\u003e\n \u003cp\u003e96.90%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"27.551020408163264%\" valign=\"top\"\u003e\n \u003cp\u003eNegative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eF1 Score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.387755102040817%\" valign=\"top\"\u003e\n \u003cp\u003e52.03%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"38.775510204081634%\" valign=\"top\"\u003e\n \u003cp\u003e39.31%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"27.551020408163264%\" valign=\"top\"\u003e\n \u003cp\u003eNeutral\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.285714285714286%\" valign=\"top\"\u003e\n \u003cp\u003eF1 Score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.387755102040817%\" valign=\"top\"\u003e\n \u003cp\u003e60.47%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"38.775510204081634%\" valign=\"top\"\u003e\n \u003cp\u003e72.71%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"V. DISCUSSION","content":"\u003cp\u003e1.\u0026nbsp; \u0026nbsp;Model Performance Evaluation\u003c/p\u003e\n\u003cp\u003eThe performance of the Naive Bayes and Gradient Boosted Machines (GBM) models in sentiment classification reveals their respective strengths and limitations. Consistent with findings from previous research\u0026nbsp;[21], the Naive Bayes model excelled in identifying positive sentiments with high precision and recall, indicative of its robust pattern recognition capabilities for positively connotated words\u0026nbsp;[10]. However, it showed limitations in classifying negative and neutral sentiments, echoing challenges noted in other sentiment analyses using Naive Bayes\u0026nbsp;[11], [12]. This model’s difficulty with these categories could be due to its probabilistic nature, which may not capture the contextual nuances essential in these more complex sentiment classifications\u0026nbsp;[22].\u003c/p\u003e\n\u003cp\u003eConversely, the GBM model displayed a commendable balance across all sentiment categories, particularly excelling in identifying neutral sentiments\u0026nbsp;[23]. This capability suggests that GBM is adept at handling the ambiguity and subtlety often associated with neutral expressions, potentially due to its iterative refinement process that enhances accuracy over successive models\u0026nbsp;[15], [16]. Despite a slight decline in precision for positive sentiments compared to Naive Bayes, GBM’s performance remained robust, showcasing its utility in applications requiring nuanced sentiment detection\u0026nbsp;[14].\u003c/p\u003e\n\u003cp\u003e2.\u0026nbsp; \u0026nbsp;Methodological Analysis\u003c/p\u003e\n\u003cp\u003eThe application of TF-IDF for data preprocessing has been validated as an effective method for preparing text for analysis\u0026nbsp;[13], [17]. However, the challenge of managing a large number of attributes (11,510), as noted in our findings, is consistent with other studies indicating potential overfitting issues in Naive Bayes when faced with high-dimensional data\u0026nbsp;[18]. Future research might explore dimensionality reduction strategies, such as those discussed by Dang et al.\u0026nbsp;[15], to enhance model performance without losing critical information.\u003c/p\u003e\n\u003cp\u003e3.\u0026nbsp; \u0026nbsp;Influence and Implications of Research Results\u003c/p\u003e\n\u003cp\u003eOur research contributes valuable insights into public sentiments regarding judicial decisions, aligning with the broader discourse on political communications and public opinion\u0026nbsp;[2], [3], [4]\u0026nbsp; The ability of the models, particularly Naive Bayes, to effectively classify positive sentiments aligns with their potential utility in monitoring public reactions to governmental decisions, a critical aspect of democratic governance\u0026nbsp;[6], [9]. However, the noted deficiencies in identifying negative and neutral sentiments highlight the ongoing need for methodological advancements to refine sentiment analysis techniques, ensuring a comprehensive understanding of public opinion\u0026nbsp;[19], [20].\u003c/p\u003e\n\u003cp\u003eThis study underscores the importance of continuous improvement in sentiment analysis methodologies to achieve balanced and accurate insights, which are essential for policymakers, social scientists, and political analysts aiming to gauge public opinion and adapt strategies accordingly [24].\u003c/p\u003e"},{"header":"VI. CONCLUSION","content":"\u003cp\u003eThis study has effectively demonstrated the use of Naive Bayes and Gradient Boosted Machines (GBM) models in performing sentiment analysis of Twitter data related to the Constitutional Court’s decision on the 2024 Indonesian elections. Our findings underscore the Naive Bayes model’s strong performance in accurately identifying positive sentiments, affirming its utility in applications where positive sentiment detection is crucial. However, the model exhibited limitations in accurately classifying negative and neutral sentiments, which underscores the necessity for enhancements in handling complex sentiment classifications.\u003c/p\u003e\n\u003cp\u003eThe GBM model, on the other hand, showed a more balanced performance across all sentiment categories. It was particularly adept at recognizing neutral sentiments, highlighting its capacity to manage the subtleties and ambiguities of such expressions. This balanced performance advocates for the use of GBM in scenarios that require nuanced sentiment detection across a broader spectrum.\u003c/p\u003e\n\u003cp\u003eMethodologically, the application of the TF-IDF technique proved to be effective for text preprocessing, though the challenge of handling a large number of attributes suggests that further improvements could be made through dimensionality reduction techniques. Such adjustments could potentially enhance the generalizability and efficiency of the models without compromising the depth of analysis.\u003c/p\u003e\n\u003cp\u003eThe implications of our study are significant for stakeholders interested in public opinion analysis, including policymakers, political analysts, and social media strategists. By improving sentiment analysis techniques, a more comprehensive and accurate understanding of public sentiments can be achieved, which is pivotal for informed decision-making.\u003c/p\u003e\n\u003cp\u003eFuture work should focus on refining these models to enhance their ability to classify negative and neutral sentiments more accurately. Additionally, exploring other machine learning techniques and incorporating more sophisticated natural language processing tools could further improve the robustness and accuracy of sentiment analysis in political and social contexts.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eThe authors declare that they have no competing interests, either financial or non-financial, that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFUNDING\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAKNOWLEDGEMENT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to express our gratitude to everyone who contributed to the successful completion of this research. Special thanks go to the faculty and staff of the Department of Computer Science at Institut Mahardika, whose insights and expertise have been invaluable throughout this project. We are particularly grateful for the guidance and support provided by Dr. Christina Juliane, whose expertise in the field of data science significantly enhanced our analytical approaches.\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. However, the resources and facilities provided by Institut Mahardika have been instrumental in carrying out this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contribution\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIndra Surya Permana and Fardhoni led the main research activities, including conceptualization, methodology design, data collection, analysis, and initial manuscript drafting. Indra Surya Permana, serving as the corresponding author, additionally handled the project administration, supervision, and prepared the final version of the manuscript, integrating contributions and feedback. Fardhoni contributed specifically to data curation and provided substantial revisions to the manuscript drafts. Christina Juliane supervised the overall project, providing critical academic guidance and intellectual input, and reviewed the manuscript to ensure the accuracy and integrity of the work. All authors read and approved the final manuscript, agreeing to be accountable for all aspects of the work.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAfnira E, Optimalisasi Media Sosial sebagai Sarana Publikasi Pengawasan Pemilu (2024) : Kasus Bawaslu Kota Tanjungpinang, \u003cem\u003eJ. Mhs. Komun. Cantrik\u003c/em\u003e, vol. 3, no. 1, Art. no. 1, May 2023, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.20885/cantrik.vol3.iss1.art4\u003c/span\u003e\u003cspan address=\"10.20885/cantrik.vol3.iss1.art4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIndonesia BBCN, Putusan MK Reaksi setelah MK tolak gugatan hasil Pilpres 2024, dari Prabowo hingga pengunjuk rasa, BBC News Indonesia. Accessed: Apr. 23, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.bbc.com/indonesia/articles/ce437lrlzwlo\u003c/span\u003e\u003cspan address=\"https://www.bbc.com/indonesia/articles/ce437lrlzwlo\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHutabarat J Menjinakkan Polarisasi dan Politik Identitas Jelang Pemilu 2024, tirto.id. Accessed: Apr. 23, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tirto.id/menjinakkan-polarisasi-dan-politik-identitas-jelang-pemilu-2024-gtQ8\u003c/span\u003e\u003cspan address=\"https://tirto.id/menjinakkan-polarisasi-dan-politik-identitas-jelang-pemilu-2024-gtQ8\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNur MF Potensi Polarisasi \u0026amp; Mencegah Konflik Pendukung di Pemilu 2024, tirto.id. Accessed: Apr. 23, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tirto.id/potensi-polarisasi-mencegah-konflik-pendukung-di-pemilu-2024-gRPS\u003c/span\u003e\u003cspan address=\"https://tirto.id/potensi-polarisasi-mencegah-konflik-pendukung-di-pemilu-2024-gRPS\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIndonesia KPU Daftar Sengketa Pemilu Presiden \u0026amp; Wakil Presiden Ri 2024: Daftar Daerah Yang Mengalami Pemungutan Dan Penghitungan Suara Ulang/Susulan/Lanjutan (Baik Seluruhnya Maupun Sebagian), Komisi Pemilihan Umum. Accessed: Apr. 23, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pemilu2024.kpu.go.id/pilpres/daftar-sengketa\u003c/span\u003e\u003cspan address=\"https://pemilu2024.kpu.go.id/pilpres/daftar-sengketa\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDhiman A, Toshniwal D (Oct. 2022) AI-based Twitter framework for assessing the involvement of government schemes in electoral campaigns. Expert Syst Appl 203:117338. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.eswa.2022.117338\u003c/span\u003e\u003cspan address=\"10.1016/j.eswa.2022.117338\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDevi WR, Chingangbam C (2021) Sentiment Analysis for Electoral Prediction Using Twitter Data, in \u003cem\u003eEmerging Technologies in Data Mining and Information Security\u003c/em\u003e, vol. 1300, A. E. Hassanien, S. Bhattacharyya, S. Chakrabati, A. Bhattacharya, and S. Dutta, Eds., in Advances in Intelligent Systems and Computing, vol. 1300., Singapore: Springer Nature Singapore, pp. 25\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/978-981-33-4367-2_3\u003c/span\u003e\u003cspan address=\"10.1007/978-981-33-4367-2_3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVindua R, Zailani AU (2023) Analisis Sentimen Pemilu Indonesia Tahun 2024 Dari Media Sosial Twitter Menggunakan Python, \u003cem\u003eJur. Ris. Kom.\u003c/em\u003e, vol. 10, no. 2, p. 479, Apr. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.30865/jurikom.v10i2.5945\u003c/span\u003e\u003cspan address=\"10.30865/jurikom.v10i2.5945\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeptian L, Aljauza T, Juliane C (2023) Analisis Sentimen Putusan Mahkamah Konstitusi terhadap Batas Usia Capres dan Cawapres Menggunakan IndoBERT, \u003cem\u003eIndonesian Journal of Computer Science\u003c/em\u003e, vol. 12, no. 6, Art. no. 6, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.33022/ijcs.v12i6.3614\u003c/span\u003e\u003cspan address=\"10.33022/ijcs.v12i6.3614\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIsnain AR, Marga NS, Alita D, Sentiment Analysis Of Government Policy On Corona Case Using Naive Bayes Algorithm, (Jan. 2021) Indonesian J Comput Cybern Syst 15(1):55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.22146/ijccs.60718\u003c/span\u003e\u003cspan address=\"10.22146/ijccs.60718\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZahra DF (2023) Analisis Sentimen penggunaan Mypertamina untuk Pembelian BBM Bersubsidi mengggunakan Algoritma Naive Bayes, \u003cem\u003eExplore. jurnal. sistem. inf. dan. telematika\u003c/em\u003e, vol. 14, no. 1, p. 94, Jul. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.36448/jsit.v14i1.3098\u003c/span\u003e\u003cspan address=\"10.36448/jsit.v14i1.3098\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAnsari S, Hans K, Khatri SK (2017) A Naive Bayes classifier approach for detecting hypervisor attacks in virtual machines, \u003cem\u003e2nd International Conference on Telecommunication and Networks, TEL-NET\u003c/em\u003e vol. 2018-January, pp. 1\u0026ndash;6, Apr. 2018, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/TEL-NET.2017.8343551\u003c/span\u003e\u003cspan address=\"10.1109/TEL-NET.2017.8343551\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuan HK, Vasarhelyi MA, Codesso M, Alzamil Z (2023) Enhancing the government accounting information systems using social media information: An application of text mining and machine learning, \u003cem\u003eInternational Journal of Accounting Information Systems\u003c/em\u003e, vol. 48, p. 100600, Mar. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.accinf.2022.100600\u003c/span\u003e\u003cspan address=\"10.1016/j.accinf.2022.100600\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNair A, Paralkar C, Pandya J, Chopra Y, Krishnan D (2021) Comparative Review on Sentiment analysis-based Recommendation system, in \u003cem\u003e6th International Conference for Convergence in Technology (I2CT)\u003c/em\u003e, Apr. 2021, pp. 1\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/I2CT51068.2021.9418222\u003c/span\u003e\u003cspan address=\"10.1109/I2CT51068.2021.9418222\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDang NC, Moreno-Garc\u0026iacute;a MN, De la Prieta F (2020) Sentiment Analysis Based on Deep Learning: A Comparative Study, \u003cem\u003eElectronics\u003c/em\u003e, vol. 9, no. 3, Art. no. 3, Mar. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/electronics9030483\u003c/span\u003e\u003cspan address=\"10.3390/electronics9030483\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMehmood K, Essam D, Shafi K, Malik MK (2019) Sentiment Analysis for a Resource Poor Language\u0026mdash;Roman Urdu, \u003cem\u003eACM Trans. Asian Low-Resour. Lang. Inf. Process.\u003c/em\u003e, vol. 19, no. 1, p. 10:1\u0026ndash;10:15, Aug. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1145/3329709\u003c/span\u003e\u003cspan address=\"10.1145/3329709\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGogia E (Dec. 2021) A Review on Sentiment Analysis techniques and applications. GRET 24\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.47893/GRET.2021.1041\u003c/span\u003e\u003cspan address=\"10.47893/GRET.2021.1041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuckartz U (2014) Qualitative text analysis: a guide to methods, practice \u0026amp; using software. Sage, Los Angeles London New Delhi Singapore Washington, DC\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMohan Kumar AV, Nandkumar AN (May 2020) A Survey on Challenges and Research Opportunities in Opinion Mining. SN COMPUT SCI 1(3):171. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s42979-020-00149-4\u003c/span\u003e\u003cspan address=\"10.1007/s42979-020-00149-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOsmani A, Mohasefi JB (2023) Opinion Mining Using Enriched Joint Sentiment-Topic Model, \u003cem\u003eInt. J. Info. Tech. Dec. Mak.\u003c/em\u003e, vol. 22, no. 01, pp. 313\u0026ndash;375, Jan. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1142/S0219622022500584\u003c/span\u003e\u003cspan address=\"10.1142/S0219622022500584\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoster J, Wagner J (2021) Naive Bayes versus BERT: Jupyter notebook assignments for an introductory NLP course, in \u003cem\u003eProceedings of the Fifth Workshop on Teaching NLP\u003c/em\u003e, Online: Association for Computational Linguistics, pp. 112\u0026ndash;114. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18653/v1/2021.teachingnlp-1.20\u003c/span\u003e\u003cspan address=\"10.18653/v1/2021.teachingnlp-1.20\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlzamzami F, Hoda M, Saddik AE (2020) Light Gradient Boosting Machine for General Sentiment Classification on Short Texts: A Comparative Evaluation. IEEE Access 8:101840\u0026ndash;101858. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/ACCESS.2020.2997330\u003c/span\u003e\u003cspan address=\"10.1109/ACCESS.2020.2997330\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhalid M, Ashraf I, Mehmood A, Ullah S, Ahmad M, Choi GS (2020) GBSVM: Sentiment Classification from Unstructured Reviews Using Ensemble Classifier, \u003cem\u003eApplied Sciences\u003c/em\u003e, vol. 10, no. 8, p. 2788, Apr. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/app10082788\u003c/span\u003e\u003cspan address=\"10.3390/app10082788\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMakhi A, Fauzan AC, SISTEM INFORMASI MANAJEMEN ADMINISTRASI PADA, Apr. LEMBAGA SERTIFIKASI PROFESI P1 UNIVERSITAS NAHDLATUL ULAMA BLITAR MENGGUNAKAN MODEL RAPID APPLICATION DEVELOPMENT (2022), \u003cem\u003ejatim\u003c/em\u003e, vol. 3, no. 1, pp. 21\u0026ndash;34, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.31102/jatim.v3i1.1423\u003c/span\u003e\u003cspan address=\"10.31102/jatim.v3i1.1423\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Sentiment Analysis, Naive Bayes, Gradient Boosted Machines, Twitter, Indonesian Election","lastPublishedDoi":"10.21203/rs.3.rs-4482093/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4482093/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis research investigates the sentiment analysis of public reactions on Twitter to the Constitutional Court\u0026rsquo;s decision regarding the 2024 Indonesian election. The study focuses on evaluating the effectiveness of Naive Bayes and Gradient Boosted Machines (GBM) in categorizing Twitter sentiments into positive, negative, or neutral. Utilizing TF-IDF vectorization to process the data, our analysis aimed to discern which model more accurately captures the nuances of public sentiment. The results indicate that while Naive Bayes shows high precision and recall in detecting positive sentiments, it performs less effectively for negative and neutral sentiments. In contrast, GBM offers a more uniform performance across all sentiment categories, with particularly strong detection capabilities for neutral sentiments. This comparative analysis underscores the strengths and limitations of each model, providing valuable insights for selecting appropriate sentiment analysis tools depending on the specific nature of the sentiment being analyzed. This study contributes to the strategic application of sentiment analysis models in monitoring and interpreting public opinions in politically significant contexts.\u003c/p\u003e","manuscriptTitle":"Public Response to the Constitutional Court’s Decision on Indonesia’s 2024 Elections","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-06-07 05:17:34","doi":"10.21203/rs.3.rs-4482093/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4891d208-24ed-4dae-aabb-61ee49eb36d8","owner":[],"postedDate":"June 7th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-06-07T05:17:36+00:00","versionOfRecord":[],"versionCreatedAt":"2024-06-07 05:17:34","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4482093","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4482093","identity":"rs-4482093","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}