Prediction of market behavior for short term stock prices using regression techniques

Abstract

Stock price Prediction always been a desired area for many institutions of finance. As price prediction in finance has long been a challenging task due to volume and speed of the data, investors are always looking for good algorithm to know the future price. The various machine learning algorithms (MLR, SVM, Random Forest etc.) used to predict and make further decision on stock market. The errors of predicted prices may be minimized, if the labeled dataset is mined in a efficient way. As the technical analysis always plays a major role to put profit in a investors pocket, a very simple algorithm is proposed for short term closing price prediction after analyzing similar types of movements of last few days prices to the historical data of that stock. A novel approach using Correlation Coefficients, Euclidian Distance and machine learning techniques is proposed to forecast a meaningful price based on the SBI data, fetched from the Yahoo Finance.