Novel Design of a Sentiment Based Stock Market Index Forecasting System

Abstract This article proposes a novel idea for creating a sentiment-based stock market index forecasting model by amalgamating price and sentiment data. The design guidelines are provided using which researchers can explore and create prediction or expert systems that can help in profitably trading the stock markets. A modular design approach is proposed for developing the prediction system. The article demonstrates steps using which the prediction system model can be built. A prototype of the model is also been developed whose performance results are provided. It is observed that there is an ample avenue for improvements that can done by the researchers for tuning up the performance of the proposed model. The model performs a unique modification to the traditional tf-idf technique and converts it into a forecasting tool. The model is simple and easy to implement with very nominal memory requirements, compared to other types of models. Nifty-50 index values were used to analyze the performance of the proposed model. This model is being developed for research purposes only.