search

mjuchli / ctc-executioner

Master Thesis: Limit order placement with Reinforcement Learning
176 stars 83 forks source link

[RL] Build deep RL model #5

Closed mjuchli closed 6 years ago

mjuchli commented 6 years ago

The idea is to lay out a DQL setup which allows to train a neural network and then predict limit level for a given state.

https://ai.intel.com/demystifying-deep-reinforcement-learning/ https://medium.freecodecamp.org/deep-reinforcement-learning-where-to-start-291fb0058c01 http://karpathy.github.io/2016/05/31/rl/ http://neuro.cs.ut.ee/demystifying-deep-reinforcement-learning/ https://keon.io/deep-q-learning/ https://github.com/farizrahman4u/qlearning4k https://keon.io/deep-q-learning/#Implementing-Mini-Deep-Q-Network-DQN https://yanpanlau.github.io/2016/07/10/FlappyBird-Keras.html (good math explanation)

mjuchli commented 6 years ago

Action prediction is currently always 0. To be investigated.