In this paper we propose fuzzyclustering algorithms for directional data. It is a new type of classification maximum likelihood procedure for mixtures of von Mises distributions. These iterative clustering algorithms...
详细信息
In this paper we propose fuzzyclustering algorithms for directional data. It is a new type of classification maximum likelihood procedure for mixtures of von Mises distributions. These iterative clustering algorithms give us a new method for analysis of grouped directional data in the plane. The procedure which embeds the fuzzyc-partitions in the model of mixtures of von Mises regressions is derived. This is used as analysis of mixtures of directional regression models. Some numerical examples are given. (c) 1997 Published by Elsevier Science B.V.
暂无评论