Multiple Regression Prediction of Stock Intraday Prices

In the global merchandise market, stock prices are highly volatile and strongly influenced by factors that are uncertain and indeterministic in nature. Hence, the stocks are traded with high risk by the individuals who fail to compete with the institutional investors thereby increasing the price and market volatility. To avoid the risk and to gain profit, chart tools and statistical techniques have been used. However, the techniques used by the statistician are univariate or bivariate in nature, and it neglects the consideration of multiple variables as a whole. Hence, a multiple regression analysis of variables affecting the stock prices are done in this paper using the transparent data available from National Stock Exchange (NSE). The significant factors are identified, stock price is predicted, and the confidence levels of responses are determined. The proposed methodology produced better prediction when compared with unassisted amateur traders.