DeepRL-TradeBot

General

Research project on applying deep reinforcement learning to perform financial market predictions. Independent research conducted by Mao L. and Songlin L., guided by professor Sicun Gao, UC San Diego Jacobs School of Engineering.

Dataset

Real-time and historical data on stocks, provided by Alpha Vintage.

Method

Deep reinforcement learning

Installation

First start a python (Python3.7 & Pip3) virtual environment as following:

python3 -m venv venv
source venv/bin/activate

Then install all neccessary packages using:

pip install -r requirements.txt

Put your own Alpha Vantage API key in the .env file under /config. To fetch data from the AlphaVantage and output to a csv file under folder data, run the following:

python3 fetchdata.py
Please enter the ticker of your stock of choice: [StockName]

Reference

• Market Making via Reinforcement Learning
• Constrained Policy Optimization