A Deep Learning based Approach to Stock Market Price Prediction using Technical indicators

Prediction of stock market data is difficult because of its complex and highly volatile nature. In this work the historical data as well as the technical indicators are implemented for the purpose of prediction. Different features are extracted using the CNN technique and further the prediction is performed using the dropout based LSTM technique. The basic aim of this study is optimization of the prediction accuracy of the stock price. Different technical indicators and historical data are taken as input data. The sub max layer is substituted with KELM (Kernel Based Extreme Learning Machine). This paper shows a CNN based hybrid system applied on a variety of sources comprising of different stock market. Various matrices are used for observing the accurateness of the proposed model. Two different stock market data are considered for this purpose. The extracted features shows more accurate result. Further it is observed that the proposed model outrun different other methods discussed in this paper