Financial market prediction using hybridized neural approach

Stock prediction is one of the ultimate garnering branches within finance which has captivated researchers across the globe. It includes classical theorems to predict market behavior. These techniques can be used substantially in the financial markets to aid investors in making a decision that is qualitatively refined. But as the world is advancing, accuracy is a matter of concern now. Legacy tools used statistical regression as a method of calculation for better performance. Back propagation techniques in combination to artificial neural networks is trending these days. However in the application of ANN for predicting the financial market, variables used for technical analysis for stock prediction is top-tier. In this paper, we present a blended approach which is a fusion of the statistical regression method and ANN algorithm to predict stock market in order to enhance reliability of predictive tools. We will use past 5 years to train our neurons. In the proposed approach, the results obtained show magnificent improvisation in getting the normalized value to be fed to neurons to train them. It will be followed by feeding the obtained output as input to the linear regression technique to calculate the predictive results. The prediction through the new approach would be satisfactorily adequate for traders and investors thus helping them to take appropriate decisions.