Quantify

This is full-stack web application that leverages a Recurrent Neural Network (LSTM model) to forecast future stock prices. While it is difficult to forecast stock prices in order to earn arbitrage returns, this project was created so that I could learn some of the basics of machine learning.

Technologies Used

• Python
• Django
• JavaScript
• Vue.js
• Keras
• Tensorflow

Set-up

Front-end

1. Change into the quantify-ui directory and install dependencies.

cd quantify-ui

npm install

2. Run the front-end application on Google Chrome.

npm run dev

Back-end

1. Create a virtual environment to run the Django application in.

py -m venv stockPredictorEnv

2. Activate the virtual environment.

stockPredictorEnv\Scripts\activate

3. Use the terminal you activated the virtual environment, run the Django application.

cd stock_forecaster

py manage.py runserver

Acknowledgments

1. For valid tickers file: https://github.com/ahnazary/Finance/blob/master/finance/src/database/valid_tickers.csv
2. https://www.youtube.com/watch?v=CbTU92pbDKw&t.
3. https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/