“FIN Bi-month Report”版本间的差异
来自cslt Wiki
| 第1行： | 第1行： | ||
Bi-month work | Bi-month work | ||
* write a unfinished paper. (a week) | * write a unfinished paper. (a week) | ||
| − | * read a phd thesis "Application of Hybrid Neural Networks and<br /> Genetic Programming in Financial Forecasting" and then give a talk. (a week) | + | * read a phd thesis "Application of Hybrid Neural Networks and <br /> Genetic Programming in Financial Forecasting" and then give a talk. (a week) |
* read some papers about reinforcement learning. (a week) | * read some papers about reinforcement learning. (a week) | ||
* develop DQNStock version1.0. (a week) | * develop DQNStock version1.0. (a week) | ||
* analysis DQNStock with quantdigger. (a week) | * analysis DQNStock with quantdigger. (a week) | ||
| − | * read some papers about continuous control and develop DQNStock version2.0 | + | * read some papers about continuous control and develop DQNStock version2.0 <br /> which implements the continuous control, but it can not converge. (a week) |
| − | + | ||
2016年8月31日 (三) 00:59的版本
Bi-month work
- write a unfinished paper. (a week)
- read a phd thesis "Application of Hybrid Neural Networks and
Genetic Programming in Financial Forecasting" and then give a talk. (a week) - read some papers about reinforcement learning. (a week)
- develop DQNStock version1.0. (a week)
- analysis DQNStock with quantdigger. (a week)
- read some papers about continuous control and develop DQNStock version2.0
which implements the continuous control, but it can not converge. (a week)
