Presentation | 1994/9/21 Robust clustering based on a maximum likelihood method for estimation of the suitable mumber of clusters Naoyuki Ichimura, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In the case of gathering a set of patterns for recognition,a prioti information about a number of clusters in the set often cannot be obtained.In addition,the set may contain the outliers. The goal of this paper is to estimate the suitable number of clusters from such a set.If a model that describes a distribution of patterns is assumed,the number of clusters can be estimated by optimizing the number of parameters in the model;a maximum likelihood method and the AIC or the MDL can be used.But,if some outliers exist,they may badly affect the estimation of the number of clusters.In this paper,to reduce the effect of the outliers, robust.clustering algorithms based on the maximum likelihood method using a ″multivariate mixture normal distribution model″ a re proposed.We call those:″Maximlim likelihood Robust Clustering″ (MARC)algorithms.The numerical experiments have been done to consider the behavior of the MARC. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | robust clustering / outliers / multivariate mixture normal distrubution model / AIC / MDL |
Paper # | PRU94-20 |
Date of Issue |
Conference Information | |
Committee | PRU |
---|---|
Conference Date | 1994/9/21(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 | Pattern Recognition and Understanding (PRU) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Robust clustering based on a maximum likelihood method for estimation of the suitable mumber of clusters |
Sub Title (in English) | |
Keyword(1) | robust clustering |
Keyword(2) | outliers |
Keyword(3) | multivariate mixture normal distrubution model |
Keyword(4) | AIC |
Keyword(5) | MDL |
1st Author's Name | Naoyuki Ichimura |
1st Author's Affiliation | Electrotechnical Laboratory() |
Date | 1994/9/21 |
Paper # | PRU94-20 |
Volume (vol) | vol.94 |
Number (no) | 241 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |