图 书 馆
提高半监督聚类：特征投影角度Enhancing Semi-Supervised Clustering: A Feature Projection Perspective |
|
| 课程网址: | http://videolectures.net/kdd07_tang_essc/ |
| 主讲教师: | Wei Tang |
| 开课单位: | 佛罗里达大西洋大学 |
| 开课时间: | 2007-12-14 |
| 课程语种: | 英语 |
| 中文简介: |  |
| 课程简介: | Semi supervised clustering uses limited supervised labeled instances or paired instance constraints to help unsupervised clustering, which often significantly improves the performance of clustering. Although expert knowledge costs a lot on this problem, most of the existing work is not suitable for dealing with high-dimensional sparse data. This paper fills this key void based on the development of a semi supervised clustering method through feature projection of spherical text (screen). Specifically, our constraint guided feature projection problem can be well integrated with semi supervised clustering algorithm, which can effectively reduce the dimension of data. In fact, the experimental results on several real data sets show that the screen method can effectively deal with high-dimensional data and provide an attractive clustering performance. |
| 关 键 词: | 半监督聚类; 高维稀疏数据; 维数 |
| 课程来源: | 视频讲座网 |
| 最后编审: | 2021-08-27：zyk |
| 阅读次数: | 56 |