Loading...
Please wait, while we are loading the content...
Knowledge discovery from local principal components independent of arbitrary factors ∗.
| Content Provider | CiteSeerX |
|---|---|
| Author | Oh, Chi-Hyon Honda, Katsuhiro Ichihashi, Hidetomo |
| Abstract | In this paper, we propose a technique of extracting local principal components independent of arbitrary factors chosen. The proposed method takes advantage of Fuzzy c-Regression Models (FCRM) to estimate the parameters of regression models for fuzzy clusters. We decompose the fuzzy scatter matrix of each cluster into two matrices by using the partial regression coefficient matrix obtained by the FCRM. One is closely related to the arbitrary factors and the other is independent of them. Solving the eigen-value problem of the decomposed matrix enables us to extract the local principal components in which influences of arbitrary factors are neutralized. We apply our method to a POS transaction data set in order to discover useful knowledge from it. 1 |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Arbitrary Factor Local Principal Component Independent Knowledge Discovery Local Principal Component Eigen-value Problem Fuzzy C-regression Model Useful Knowledge Partial Regression Coefficient Matrix Decomposed Matrix Po Transaction Data Set Fuzzy Scatter Matrix Regression Model Fuzzy Cluster |
| Content Type | Text |
