Multivariate Stock Price Prediction Model

It was part of final year ML course project. This project have two sub-systems:
Data Collection: Mined stock related trade information since January, 2012 to June, 2018 from Dhaka Stock Exchange website. We later installed an auto update feature to it to continuously monitor price changes within a day.
Prediction Model: Designed and deployed a RNN with LSTM model for predicting stock price trends. Although as a learning project it were successful, we would highly discourage anyone to invest following any such ML based prediction.
We are also skeptical about any large scale accurate stock price predictor as the stock market resembles an type 2 chaotic system that reacts to the prediction itself.

Applied Technologies:

• Tensorflow & Keras: For model Development
• Selenium (Java) & JSoup: For automated data capture, parsing & building dataset.