EcoSta 2017: Start Registration
View Submission - EcoSta2017
A0693
Title: The multimix class of mixture models for clustering mixed categorical and continuous data Authors:  Lynette Hunt - University of Waikato (New Zealand) [presenting]
Murray Jorgensen - Auckland University of Technology (New Zealand)
Abstract: One possible approach to clustering data is to assume that the data to be clustered come from a finite mixture of populations. The mixture likelihood approach has been well developed and is much used, especially for mixtures where the component distributions are multivariate normal. However, when clustering large multivariate data sets, it is rare to find data with all attributes being either continuous or categorical. The Multimix class of mixture models was proposed in the nineties. This approach enables the clustering of mixed categorical and continuous data, and can also cope with mixed data where data are missing at random in the sense. We demonstrate the Multimix approach to clustering mixed categorical and continuous data, and illustrate the methodology by clustering a large data set.