作者: Robert Burduk , Pawel Trajdos
DOI: 10.1007/978-3-642-40846-5_14
关键词:
摘要: This paper presents the problem of building sequential model classification task. In our approach structure is built in learning phase classification. this a split criterion based on broken stick proposed. The distribution created for each column confusion matrix. associated with analysis received distributions. obtained results were verified ten data sets. Nine sets come from UCI repository and one real-life set.