Model-based clustering via mixture of skew-t distribution with missing information
碩士 === 國立中興大學 === 統計學研究所 === 102 === Multivariate mixture modeling approach using the skew-t distribution has been recently examined as a powerful and flexible tool for robust model-based clustering and classification. Missing data are a ubiquitous problem for researchers encountered in practice. In...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2014
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/80760298934975499687 |
| Summary: | 碩士 === 國立中興大學 === 統計學研究所 === 102 === Multivariate mixture modeling approach using the skew-t distribution has been recently examined as a powerful and flexible tool for robust model-based clustering and classification. Missing data are a ubiquitous problem for researchers encountered in practice. In this thesis, we offer a computationally flexible EM-type procedure for maximum likelihood estimation of multivariate skew-t mixture models when missing values occur in data. Further, we present a common information-based approach to approximating the asymptotic covariance matrix of the ML estimator using the outer product of the scores. To assist the development and ease the implementation of our algorithm, we make use of two auxiliary permutation matrices for fast determining
the observed and missing parts of each observation. The practical usefulness of the proposed methodology is illustrated through simulations with varying proportions of artificial missing values and a real data example with genuine missing values.
|
|---|