“FIN Status Report 2016-11-28”版本间的差异

2016年11月28日 (一) 08:19的最后版本

People	Last Week	This Week
Yang Wang	check drrl and dqt. confirm that the fitted model of drrl can make profit.	establish a backtesting framework do backtesting to the models.
Yuxin Zhang	Do comparable experiments to confirm the efficiency of using 1min data. Add volume data and bid-ask data as input for dqt. Share a paper about transfer learning.	Implement multi-layer-perceptron algorithm. Explore other research branches (Multi-factor, correlation of stocks).

@@ 第13行： / 第13行： @@
 |-
 | Yuxin Zhang||
-Use 1min daily data for dqt and acquire good results.<br>
+Do comparable experiments to confirm the efficiency of using 1min data.<br>
-Design automated network establishing strategy.<br>
+Add volume data and bid-ask data as input for dqt.<br>
-Read some papers about local minima in training neural networks.<br>
+Share a paper about transfer learning.<br>
 ||
-Try to add factors as part of the input data.<br>
+Implement multi-layer-perceptron algorithm.<br>
-Implement experiments of drrl using 1min data.
+Explore other research branches (Multi-factor, correlation of stocks).
 |}