Run CNN and DQN model with Tensorflow for stock prediction.
Feature include daily close price, daily relative price, MA, RSI.
Stock data are collected from matplotlib.finance.
- Tensorflow
- Python 2.7
- matplotlib
- Numpy
- Gym (for Deep Q Network)
Distributed version for MNIST example.
Create a cluster of Tensorflow servers, and distribute a computation graph across the cluster.
Need to consider about transmission time. Transmission time must not be longer than compute time.
Take reference from RobRomijnders's work (http://robromijnders.github.io/CNN_tsc/)
Take reference from
- FloodSung's work (https://zhuanlan.zhihu.com/p/21477488?refer=intelligentunit)
- yenchenlin's work (https://github.com/yenchenlin/DeepLearningFlappyBird)