Github项目推荐 | 股市预测的机器学习/深度学习模型/资源集锦

Stock Prediction Models - Gathers machine learning and deep learning models for Stock forecasting, included trading bots and simulations

by Husein Zolkepli

Github项目地址：

https://github.com/huseinzol05/Stock-Prediction-Models

股票预测模型，收集了机器学习和股票预测的深度学习模型，包括交易机器人和（股票）模拟。

模型

Stacking models - 堆叠模型

1. 深度前馈自动编码器神经网络降低尺寸+深度递归神经网络+ ARIMA +极限提升梯度回归

2. AdaBoost算法+Bagging 算法+ExtRa Trees极度随机树+梯度增强+随机森林+ XGB

Deep-learning models - 深度学习模型

1. LSTM Recurrent Neural Network - LSTM递归神经网络

2. Encoder-Decoder Feed-forward + LSTM Recurrent Neural Network - 编码器 - 解码器前馈+ LSTM递归神经网络

3. LSTM Bidirectional Neural Network - LSTM双向神经网络

4. 2-Path LSTM Recurrent Neural Network - 2路LSTM递归神经网络

5. GRU Recurrent Neural Network - GRU递归神经网络

6. Encoder-Decoder Feed-forward + GRU Recurrent Neural Network - 编码器 - 解码器前馈+ GRU递归神经网络

7. GRU Bidirectional Neural Network - GRU双向神经网络

8. 2-Path GRU Recurrent Neural Network - 2路GRU递归神经网络

9. Vanilla Recurrent Neural Network - 简单循环神经网络

10. Encoder-Decoder Feed-forward + Vanilla Recurrent Neural Network - 编码器 - 解码器前馈+简单循环神经网络

11. Vanilla Bidirectional Neural Network - 简单双向神经网络

12. 2-Path Vanilla Recurrent Neural Network - 双路简单循环神经网络

13. LSTM Sequence-to-Sequence Recurrent Neural Network - LSTM序列到序列递归神经网络

14. LSTM with Attention Recurrent Neural Network - 具有注意递归神经网络的LSTM

15. LSTM Sequence-to-Sequence with Attention Recurrent Neural Network - 具有注意递归神经网络的LSTM序列到序列

16. LSTM Sequence-to-Sequence Bidirectional Recurrent Neural Network - LSTM序列到序列双向递归神经网络

17. LSTM Sequence-to-Sequence with Attention Bidirectional Recurrent Neural Network - 具有注意双向递归神经网络的LSTM序列到序列

18. LSTM with Attention Scaled-Dot Recurrent Neural Network - 具有注意缩放点递归神经网络的LSTM

19. LSTM with Dilated Recurrent Neural Network - 具有扩张的递归神经网络的LSTM

20. Only Attention Neural Network - 单注意神经网络

21. Multihead Attention Neural Network - 多头注意神经网络

22. LSTM with Bahdanau Attention - 具有BaDayau注意力的LSTM

23. LSTM with Luong Attention - 具有Luong注意力的LSTM

24. LSTM with Bahdanau + Luong Attention - 具有BaDayau+Luong注意力的LSTM

25. DNC Recurrent Neural Network - DNC递归神经网络

26. Residual LSTM Recurrent Neural Network - 残余LSTM递归神经网络

27. Byte-net

28. Attention is all you need

29. Fairseq

Agents - 代理

1. Turtle-trading agent

2. Moving-average agent

3. Signal rolling agent

4. Policy-gradient agent

5. Q-learning agent

6. Evolution-strategy agent

7. Double Q-learning agent

8. Recurrent Q-learning agent

9. Double Recurrent Q-learning agent

10. Duel Q-learning agent

11. Double Duel Q-learning agent

12. Duel Recurrent Q-learning agent

13. Double Duel Recurrent Q-learning agent

14. Actor-critic agent

15. Actor-critic Duel agent

16. Actor-critic Recurrent agent

17. Actor-critic Duel Recurrent agent

18. Curiosity Q-learning agent

19. Recurrent Curiosity Q-learning agent

20. Duel Curiosity Q-learning agent

21. Neuro-evolution agent

22. Neuro-evolution with Novelty search agent

23. ABCD strategy agent

Data Explorations - 数据探索

1. 关于TESLA股票的股票市场研究，tesla-study.ipynb

2. 时尚趋势预测与交叉验证，  fashion-forecasting.ipynb

3. 使用LSTM预测进行比特币分析，bitcoin-analysis-lstm.ipynb

4. 使用k-means、SVM和高斯分布对特斯拉股票进行异常值研究  outliers.ipynb

5. Kijang Emas Bank Negara， kijang-emas-bank-negara.ipynb

Simulations - 模拟

1. 使用蒙特卡罗进行股市模拟，stock-forecasting-monte-carlo.ipynb

2. 利用蒙特卡罗-马尔可夫链模拟股票市场，mcmc-stock-market.ipynb

Tensorflow-js

我在Tensorflow JS中编写了LSTM递归神经网络和简单的信号滚动代理，你可以在 huseinhouse.com/stock-forecasting-js 中试试。

结果

Results agent（部分）

这个代理每次交易只能买入或卖出1个单位。

1.Turtle-trading agent, turtle-agent.ipynb

2.Moving-average agent, moving-average-agent.ipynb

3.Signal rolling agent, signal-rolling-agent.ipynb

4.Policy-gradient agent, policy-gradient-agent.ipynb

5.Q-learning agent, q-learning-agent.ipynb

注：由于篇幅过长，剩余的结果图请移步本项目查看>>>https://github.com/huseinzol05/Stock-Prediction-Models

Results free agent

这个代理可以在每笔交易中买卖n个单位。

进化策略代理 evolution-strategy-agent.ipynb

total gained 11037.529911, total investment 110.375299 %

贝叶斯代理的进化策略 evolution-strategy-bayesian-agent.ipynb

total gained 13295.469683, total investment 132.954697 %

结果信号预测

LSTM递归神经网络

LSTM双向神经网络

二路LSTM递归神经网络

深度前馈自动编码器神经网络降低尺寸+深度递归神经网络+ ARIMA +极限提升梯度回归

LSTM序列到序列递归神经网络

具有注意递归神经网络的LSTM序列到序列

具有注意双向递归神经网络的LSTM序列到序列

编码器 - 解码器前馈+ LSTM递归神经网络

Adaboost + Bagging + Extra Trees + Gradient Boosting + Random Forest + XGB

点击阅读原文，查看本文更多内容