Study on Integrated Machine Learning Model for Investment Decision on Indian Stocks based on Fundamental and Sentimental Analysis

Abstract

Stock market has been an active area of research, still remains challenging because of the volatile and multifaceted nature of the market. Investors usually use traditional stock data, which has limitations to formulate investment decisions. Due to the rise of Internet, and analytics platforms, new forms of information, are gaining more importance in Indian stock market. Despite, most of the investors find it a tedious task to make correct choice when it comes to profitable investments. newlineThe fundamental, sentimental and technical analysis independently addresses the problem of prediction and creation of a profitable portfolio for the investors to a certain extent. But the accuracy and reliability of these methods remained a concern for most of the traders. However it is evident from the market price movements that the sentimental overflow of the investors highly influences the fundamental prices. Integrated model are built on the interaction effect of one another. Hence the proposed method studied the significance of combining the best suited analysis for suggesting the stocks for gainful investments. Empirical analysis of the fundamental models on the data revealed the suitability of the models for the prediction of the stock prices. The use of multiple regression machine learning approach to revise and refine fundamental models improved the accuracy of the valuation models. newlineThe sentimental analysis using the VADER algorithm has been applied to the news headlines, collected from websites (moneycontrol.com, Twitter), and it was observed that the analysis had a great impact on the price movements in the market. The model identified the sentiments pertaining to a stock in the market from the news headlines and tweets. The new integrated equity valuation model (NiSa model) combining fundamental and sentimental analysis along with the random-forest classifier has been proposed to provide investment suggestions in the Stock market. The stocks are classified into buy, sell, overvalued short-term buy (OSB) and undervalue