Risk Analysis of Indian Banking Stocks Using Machine Learning Techniques

Indian banking stocks are highly volatile in nature. Although these stock values are dependent on their fundamentals. However, the banking stocks are susceptible to economic ups and downs, market sentiments, manufacturing factors and employment conditions, etc. Existing studies largely focus on implementing machine learning models for stock price prediction. However, minimal studies exist for stock risk analysis. In order to safeguard the investment of retail customers, there is a need to analyze the risk associated with them by simple tools in advance. In this paper, machine learning techniques viz. Long Short-Term Memory (LSTM) and Extreme Gradient Boosting (XGBoost) are investigated to predict the banking stocks risk. The statistical risk measures used are beta and value at risk (VaR). The simulation results indicate that the least Mean Absolute Error (MAE) for beta and Value at Risk are obtained by XGBoost and LSTM techniques respectively.