Lexscope: An Ambiguity-Aware Transformer-Based Framework For Aspect-Based Sentiment Analysis

Lexscope: An Ambiguity-Aware Transformer-Based Framework For Aspect-Based Sentiment Analysis | 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 Lexscope: An Ambiguity-Aware Transformer-Based Framework For Aspect-Based Sentiment Analysis Supipi Anuththara, Saadh Jawwadh This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9674471/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 The rapid growth of user-generated reviews across e-commerce and service platforms has increased the need for fine-grained sentiment analysis. Aspect-Based Sentiment Analy sis (ABSA) identifies sentiment toward specific aspects within a sentence, offering more detailed insights than traditional document-level approaches. However, existing Transformer-based models rely on implicit contextual representations and do not explicitly handle lexical ambiguity—where aspect terms carry multiple meanings depending on context—leading to incor rect sentiment predictions. This paper proposes LexScope, an ambiguity-aware Transformer-based framework for ABSA that fine-tunes RoBERTa-base and integrates a gloss-informed Word Sense Disambiguation (WSD) module using WordNet definitions and SBERT cosine similarity. The selected word sense is ap pended as a structured semantic hint to the model input to improve semantic understanding. WSD is applied selectively via a confidence threshold to avoid introducing noise. Additionally, LIME and sense-level explanations are incorporated to improve interpretability. Evaluated on the SemEval-2014 Task 4 dataset, LexScope achieves 88.11% accuracy and a Macro F1 of 0.8420, outperforming a strong RoBERTa baseline (85.61% accuracy, 0.7998 Macro F1). On a lexically ambiguous subset, the im provement is further pronounced, confirming that explicit sense disambiguation enhances both classification performance and interpretability in ABSA. Theoretical Computer Science Aspect-Based Sentiment Analysis Word Sense Disambiguation RoBERTa Lexical Ambiguity Explainable AI SemEval-2014 Natural Language Processing Full Text Additional Declarations The authors declare no competing interests. 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-9674471","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":637900640,"identity":"0263124b-1bca-47b3-b5c3-b6b821fedeb8","order_by":0,"name":"Supipi Anuththara","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABD0lEQVRIie2RsUrEQBCGJyzsNXOk3XAheYWRtPcwCRZpjKQSi4AbDtZGtM29hY+wYSHVYX2lV10jIthcoWDWdHcm11rsVw58zD/zAzgc/xQNBBGf1fVreltFnA3TUJxREh/Nij42XcLZ4OCUYsnWTa6CtWKZhDNK3BStKUvwnnUmF3PF88eZr+FQAQbyb4W216lpCBjpVib4EhaqD+Y9dICLkS0krsggAae2lpd4w38VmEvAaDTYoCAZTxrkLLfne98TCmwHRQTKq+tGsdQqzG4ZDbZ5t7cI8tFbQf/kiz4YmbATo+fH94X5LL+Wdyre7w99lXH81O52b9UyEnosGdoejjrQJ5NTxeFwOBwT/ACe01GLatnGCwAAAABJRU5ErkJggg==","orcid":"","institution":"Informatics Institute of Technology","correspondingAuthor":true,"prefix":"","firstName":"Supipi","middleName":"","lastName":"Anuththara","suffix":""},{"id":637900641,"identity":"e07c430c-981e-4535-92f1-af75aba3c402","order_by":1,"name":"Saadh Jawwadh","email":"","orcid":"","institution":"Informatics Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"Saadh","middleName":"","lastName":"Jawwadh","suffix":""}],"badges":[],"createdAt":"2026-05-11 04:34:29","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true},"doi":"10.21203/rs.3.rs-9674471/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9674471/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109098976,"identity":"6bc7f40c-0f2e-4590-8866-9b1965c53d2c","added_by":"auto","created_at":"2026-05-12 14:15:31","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":120003,"visible":true,"origin":"","legend":"","description":"","filename":"paper.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9674471/v1_covered_acf3a130-0bc4-4293-af9d-077b450435dd.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eLexscope: An Ambiguity-Aware Transformer-Based Framework For Aspect-Based Sentiment Analysis\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Informatics Institute of Technology","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"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":"Aspect-Based Sentiment Analysis, Word Sense Disambiguation, RoBERTa, Lexical Ambiguity, Explainable AI, SemEval-2014, Natural Language Processing","lastPublishedDoi":"10.21203/rs.3.rs-9674471/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9674471/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe rapid growth of user-generated reviews across e-commerce and service platforms has increased the need for fine-grained sentiment analysis. Aspect-Based Sentiment Analy sis (ABSA) identifies sentiment toward specific aspects within a sentence, offering more detailed insights than traditional document-level approaches. However, existing Transformer-based models rely on implicit contextual representations and do not explicitly handle lexical ambiguity—where aspect terms carry multiple meanings depending on context—leading to incor rect sentiment predictions. This paper proposes LexScope, an ambiguity-aware Transformer-based framework for ABSA that fine-tunes RoBERTa-base and integrates a gloss-informed Word Sense Disambiguation (WSD) module using WordNet definitions and SBERT cosine similarity. The selected word sense is ap pended as a structured semantic hint to the model input to improve semantic understanding. WSD is applied selectively via a confidence threshold to avoid introducing noise. Additionally, LIME and sense-level explanations are incorporated to improve interpretability. Evaluated on the SemEval-2014 Task 4 dataset, LexScope achieves 88.11% accuracy and a Macro F1 of 0.8420, outperforming a strong RoBERTa baseline (85.61% accuracy, 0.7998 Macro F1). On a lexically ambiguous subset, the im provement is further pronounced, confirming that explicit sense disambiguation enhances both classification performance and interpretability in ABSA.\u0026nbsp;\u003c/p\u003e","manuscriptTitle":"Lexscope: An Ambiguity-Aware Transformer-Based Framework For Aspect-Based Sentiment Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-12 14:08:18","doi":"10.21203/rs.3.rs-9674471/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":"5fece3f4-e1d0-4cac-967e-4953501fb54b","owner":[],"postedDate":"May 12th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":67891693,"name":"Theoretical Computer Science"}],"tags":[],"updatedAt":"2026-05-12T14:08:18+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-12 14:08:18","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9674471","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9674471","identity":"rs-9674471","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}