Stock Market Prediction Using Machine Learning: Evidence from India

Literature deciphers the dynamics of the stock market environment across the regions. Moreover, the emerging stock market like India has been experiencing several ups and downturns due to its continuous economic reforms since the early 1990s, which makes the Indian stock markets exhibit the diversified information characteristics. The chapter predicts the movements of the Indian stock markets over 2000–2022, and observes certain dynamism in both the actual and predicted trends of the Indian stock markets. The results revealed that Long Short-Term Memory holding the time-independence characteristics and greater extent of prediction accuracy proved as the best machine learning technique to predict the movement of the Indian stock markets. Moreover, the degree of prediction accuracy of all the machine learning techniques except Long-short term memory varies from one time to other. On the other hand, Support vector machines and linear regression models with their lowest degree of prediction accuracy and highest errors proved least appropriate in predicting the movements of Nifty, and Sensex respectively. The robustness of our method would benefit for testing it on another markets, and time periods. The study also discusses the strengths and weaknesses of several machine learning techniques and provide important insights in applying advanced technologies for stock market prediction of an emerging economy like India. Our prediction approach provides a potentially beneficial alternative for the investors to identify the return opportunities and achieve the diversification benefits by mitigating risk while investing in the Indian stock markets.