A Novel Approach to Improve Accuracy in Stock Price Prediction using Gradient Boosting Machines Algorithm compared with Naive Bayes Algorithm

The paper's goal is to evaluate the reliability of stock price forecasts made using stock values by Gradient Boosting Machines A as opposed to the Naive Bayes Algorithm. Sample size for the Gradient Boosting Machines (GBM) Algorithm is 20. and Naive Bayes Algorithm is iterated several times for estimating the accuracy pricing for stocks. The Gradient Boosting Machines Algorithm's Novel Loss Function, which is based on prior stock prices, helps to reduce the total prediction error. Compared to the accuracy of the Naive Bayes algorithm (87.7%), the GBM method is much more accurate (92.3%).