| Presentation | 2012-09-02 Information theoretic clustering using competitive learning Comparsion of criterion functions and algorithms for document clustering Toshio UCHIYAMA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Information-theoretic clustering (ITC) finds clusters based on the similarity of the distributions of features. An ITC algorithm based on optimizing the clustering criterion has previously been proposed. This algorithm is reminiscent of the k-means algorithm, but uses Kullback-Leibler (KL) divergence when updating the cluster-labels of the data. Recently, a novel method, based on the idea above, has been proposed. It uses competitive learning, which is known to be superior to the k-means algorithm. The method also uses skew divergence instead of KL divergence to avoid the zero-frequency problem. This paper shows that the method performs better than existing clustering algorithms, such as maximum margin clustering and a method based on mixture of von Mises-Fisher distribution, when applied to text data sets in multiclass problems. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Information-theoretic clustering / Competitive learning / Skew divergence / Kullback-Leibler divergence |
| Paper # | PRMU2012-33,IBISML2012-16 |
| Date of Issue |
| Conference Information | |
| Committee | IBISML |
|---|---|
| Conference Date | 2012/8/26(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Information-Based Induction Sciences and Machine Learning (IBISML) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Information theoretic clustering using competitive learning Comparsion of criterion functions and algorithms for document clustering |
| Sub Title (in English) | |
| Keyword(1) | Information-theoretic clustering |
| Keyword(2) | Competitive learning |
| Keyword(3) | Skew divergence |
| Keyword(4) | Kullback-Leibler divergence |
| 1st Author's Name | Toshio UCHIYAMA |
| 1st Author's Affiliation | NTT Service Evolution Laboratories() |
| Date | 2012-09-02 |
| Paper # | PRMU2012-33,IBISML2012-16 |
| Volume (vol) | vol.112 |
| Number (no) | 198 |
| Page | pp.pp.- |
| #Pages | 8 |
| Date of Issue | |