“Yang Wang 2016-04-25”版本间的差异
来自cslt Wiki
| （相同用户的一个中间修订版本未显示） | |||
| 第2行： | 第2行： | ||
* found word2vec was useless for chat disentanglement task. | * found word2vec was useless for chat disentanglement task. | ||
* read some papers about correlation clustering. | * read some papers about correlation clustering. | ||
| + | * used t-sne for clustering | ||
* downloaded a big ubuntu chat corpus . | * downloaded a big ubuntu chat corpus . | ||
To Do Next Week: | To Do Next Week: | ||
| − | * try to use new correlation clustering algorithm. | + | * try to use a new correlation clustering algorithm. |
* try to make use of the big corpus. | * try to make use of the big corpus. | ||
2016年4月25日 (一) 00:47的最后版本
Weekly Report:
- found word2vec was useless for chat disentanglement task.
- read some papers about correlation clustering.
- used t-sne for clustering
- downloaded a big ubuntu chat corpus .
To Do Next Week:
- try to use a new correlation clustering algorithm.
- try to make use of the big corpus.
