Improved Deep Learning based Model for Stock Trend Prediction using Long Short Term Memory

Abstract

Stock market strategies are highly sophisticated and rely on massive newlineamounts of data. It has been a tedious task for many experts and newlineinvestors to analyze and calculate stock prices in the future. Many newlinemachine learning techniques have been found to deal with complicated newlinecomputational problems and effi-cient predictive methods without any newlinecomplex programming. This study aims to investigate the capabilities of newlineLong Short-Term Memory, a form of Recurrent Neural Network, in newlinepredicting future stock values. The prediction of LSTM is also compared newlinewith KNN, SVM, and RNN models. This paper uses three years data of newlinedifferent companies such as Adani ports, Asian paint, Axis Bank, Cipla, newlineHcltech, Hdfc, and Titan. MSE and R2 newline measures are used to compare newlineresults. newline