wassname/rl-portfolio-management
1
0
Fork 0
mirror of https://github.com/wassname/rl-portfolio-management.git synced 2026-08-07 05:30:10 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update readme.md

Browse Source
This commit is contained in:
Mike Clark
2017-10-29 15:39:38 +08:00
committed by GitHub
parent a672367197
commit ce727cffa4
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
readme.md
+1 -1
View File
@@ -34,7 +34,7 @@ I have managed to overfit to the training data with no trading costs but it coul
- VPG model, test: 1.00 in 15 hours, 1.0 in 50 days
- [DDPG model](https://github.com/wassname/rl-portfolio-management/blob/8c74f136765f621eb45d484553b9f778e9243a84/keras-ddpg.ipynb), test
![](docs/tensorforce-VPG-test.png)
![](https://raw.githubusercontent.com/wassname/rl-portfolio-management/8c74f136765f621eb45d484553b9f778e9243a84/docs/tensorforce-VPG-test.png)
This test period is directly after the training period and it looks like the usefullness of the models learned knowledge may decay as it moves away from its training interval.