Visualising and Forecasting Stocks using XGBoost Algorithm

Investing in the stock market has always been a popular way for people to generate high returns. Nowadays, even a common man can invest in the stock market and generate good profits if he/she analyses the previous data correctly. Hence, extensive research is needed in the methodology of stock price movements to predict stock prices accurately. The literature survey of this study explored other algorithms like linear regression, pace regression, sentiment analysis, etc for the same. This paper explores using the XGBoost algorithm for the price prediction of stocks in the Indian stock market due to its better accuracy than other models. The system integrates the XGBoost algorithm to train the model on Indian stock market data. The uniqueness of this system is that it entails the use of the XGBoost algorithm, a seldom-used algorithm for predicting stock prices in the Indian stock market. The proposed system also has the distinct characteristic inclusion of the Indian stock market data which is rarely seen in many stock price prediction systems. The accuracy of the model obtained is approximately 84.42 % on average. By referencing the research literature, this study also explores other machine-learning techniques. It concludes that though the XGBoost model is highly accurate in many instances, it is best used in conjunction with different models to enhance their utility.