Stock Market PredictionWeb Service Using Deep Learning by LSTM

摘要

Predicting the future price of stocks using closing price via LSTM, an artificial recurrent neural network is presented in this paper. Here the dataset of previous stock closing prices from DSE Bangladesh has been used to train the module. Deep learning using the TensorFlow library has been used to train the system for predicting price in graphical format. The system is live as a web service using Django and android applications. Test results have given 70% accurate prediction stats. Currently, the service has a free subscription for users and will be charged in the future.