Statistical Classification

本建议考虑对控制多变量对象和变量聚类的理论进行基础统计研究，以及分类和概率在预测和解释中的相互关系。重点将放在分类概率，概率匹配，概率搜索算法，拓扑和块聚类，以及多模态。这些技术的应用包括生物实体的分组和最终识别，块投票行为的分析以及大型数据库中多模态集合的确定。

This proposal considers fundamental statistical research into the theories governing clustering of multivariate objects and variables, and the interrelation of classification and probability in prediction and explanation. The emphases will be on classification probabilities, probability matching, probability searching algorithms, topological and block clustering, and multimodality.Applications of these techniques include grouping and ultimate identification of biological entities, the analysis of block voting behavior and the determination of multimodal collections in large databases.