“FIN Status Report 2017-2-20”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
(3位用户的4个中间修订版本未显示)
第6行: 第6行:
 
give a talk about understanding deep learning.<br>
 
give a talk about understanding deep learning.<br>
 
try RMSProp optimizer, weight decay and obtain smooth predicted sequences.<br>
 
try RMSProp optimizer, weight decay and obtain smooth predicted sequences.<br>
try industry stock and obtain high correlation predicted sequence.<br>
+
try industry stock and obtain highly correlated predicted sequence.<br>
try to add index data and obtain more accuracy predicted value.
+
try to add index data and obtain more accurate predicted value.
 
||  
 
||  
test our strategy<br>
+
study the stock selecting model.<br>
1. impact of the optimizer<br>
+
study the architecture of the network.
2. impact of the clipping<br>
+
 
3. investigate the volatility problem on noisy cosine.<br>
+
4. improve our strategy.<br>
+
5.different stock set.
+
 
|-
 
|-
 
| Yuxin Zhang||  
 
| Yuxin Zhang||  
<br>
+
Try to use DDPG to predict a better limit order price.<br>
  
 
||  
 
||  
 
+
Debugging DDPG and do experiments.
 
|-
 
|-
  
 
|-
 
|-
 
| Yaodong Wang||  
 
| Yaodong Wang||  
make a view in mysql , handle undating data and choose the stable stock
+
Pick stable stocks and  Leading stocks
 
||  
 
||  
multiple factor analysis
+
Continue to pick Leading stocks and Make database smoothy  to use
 
+
 
|-
 
|-
 
| Tongzheng Ren||  
 
| Tongzheng Ren||  
Implement the strategy.
+
Implement:<br>
 +
regularization: L2-norm, dropout, batch normalization(including some bugs)<br>
 +
filters: MA, wavelet<br>
 +
Implement index data.<br>
 +
Improve strategy.<br>
 
||  
 
||  
Test our strategy.
+
Improve strategy.
 
|}
 
|}

2017年2月20日 (一) 08:08的最后版本

People Last Week This Week
Yang Wang

give a talk about understanding deep learning.
try RMSProp optimizer, weight decay and obtain smooth predicted sequences.
try industry stock and obtain highly correlated predicted sequence.
try to add index data and obtain more accurate predicted value.

study the stock selecting model.
study the architecture of the network.

Yuxin Zhang

Try to use DDPG to predict a better limit order price.

Debugging DDPG and do experiments.

Yaodong Wang

Pick stable stocks and Leading stocks

Continue to pick Leading stocks and Make database smoothy to use

Tongzheng Ren

Implement:
regularization: L2-norm, dropout, batch normalization(including some bugs)
filters: MA, wavelet
Implement index data.
Improve strategy.

Improve strategy.