Skip to content

Trading Strategy Using Keras/Tensorflow LSTM and Custom Loss Function

License

Notifications You must be signed in to change notification settings

vishalbelsare/DeepTradeRiskEsti_TXF

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 

Repository files navigation

DeepTradeRiskEsti_TXF

Trading Strategy Using Keras/Tensorflow LSTM and Custom Loss Function

使用LSTM來做台指期預測,自定義損失函數

損失函數概念參考自 https://github.com/happynoom/DeepTrade

用意在最大化收益率。

但因為原始的概念是用在操作股票,reLU激勵之後只會有正數,如果用在台指期上就變成只會做多了 因此在keras 的activations 下面自己新增了一個 relu_inverse 函數,截斷x為正的情形,變成只輸出負數 (請參考__keras_modifications__) ,修改位置位於 \Anaconda35\Lib\site-packages\keras

模型訓練: start_train.bat (可修改裡面參數,0 為model0,只輸出做多訊號,1 為 model1,只輸出做空訊號 ) 模型驗證: start_predict.bat ,因為train完模型輸出名稱一律叫做output.30,要自己改名成model0_positive 或是 model1_negative

dataset download: https://drive.google.com/open?id=0B9JaGZQ0H8ksNWNDQ1pkSzhhc1E

下載完後放在 datasetLoader 目錄下

About

Trading Strategy Using Keras/Tensorflow LSTM and Custom Loss Function

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%