Dongxu Zhang 2015-12-21
来自cslt Wiki
Last Week
- Experiments on another dataset called Webkb.
- Costs of multigram learning does not go down. Too sparse.
- training wordembedding on wiki9. Waiting for results. (power off)
- Tune svm classifier slightly improve the performance of our model but still far away from other models' performance
This Week
- deal with multigram problem.
- deal with svm classifier.
- get results from word embedding with similarity task.
