作者: Andrea Cerioli , Marco Riani , Anthony C. Atkinson
DOI: 10.1007/978-3-7908-1709-6_41
关键词:
摘要: The forward search provides a powerful and computationally simple approach for the robust analysis of multivariate data. In this paper we suggest new algorithm clustering categorical observations. Classification based on information poses number challenging issues that are addressed by our algorithm. These include selection groups, identification outliers stability suggested solution. performance is shown with both simulated real examples.