An Empirical Study on Application of Stock Price Prediction Using Deep Learning Algorithm

Stock price prediction is full of research problem in machine learning. It depends on a large number of factors which contribute to changes in the supply and demand of stock share prediction and analysis of the stock market dataset which plays a significant role in today's economy. The cycle in the financial exchange is conspicuous with part of flightiness so it is a lot of influenced by parcel numerous elements. This turned into a significant undertaking in business and money. There are assortment of calculations that are utilized for foreseeing/gauaging. They are by and large classified into two sorts. One is direct model and the other one is non-straight model. Auto Regression [AR], Auto Regression moving Average [ARMA], Auto Regression Integrated Moving Average [ARIMA] are arranged as straight models. Neural Networks are non-straight models. Fake Neural Networks and profound learning are the most overall instruments utilizing python program. Such methods square measure accommodating in learning progressed kinds of data by exploitation models of directed learning exploitation essential libraries. In profound learning, from any place a program is composed, it'll be getting customized to be advised gradually the strategy to perform insightful errands of execution on the far side the programming limits it step by step acquires through stock datasets and expertise. in view of high volumes of data getting produced accessible business sectors, machines would master separating designs, in that approach making genuinely reasonable forecasts. various Deep learning models/designs square measure effectively available for stock worth expectation of a particular organization.
- Recurrent Neural Networks(RNN)
- Convolutional Neural Networks(CNN)
- Artificial Neural Networks
LSTM model is made out of a successive information layer followed by three lstm layers and a thick layer with enactment and afterward at long last a thick yield layer with the direct actuation work. In this research project KERAS with LSTM models has been explored. Historical stock price data was obtained from yahoo finance and used to build LSTM model using python codes to visualize real and predicted price and using regression equation function to forecast stock price and calculating mean absolute error, mean squared error and mean absolute percentage error to find accuracy.