Study on Fuzzy Co-clustering of Incomplete Data

点击翻译按钮获取中文摘要

Cooccurrence information analysis is an important technique for handling such cooccurrence information as purchase history data or document-keyword frequency data with the goal of extracting co-clusters of mutually familiar pairs of objects and items. In applying co-clustering algorithms to collaborative filtering and document summarization, it is needed to handle such data incompleteness as missing elements, noise and non-Euclidean nature. In this research, several co-clustering algorithms, which can handle incomplete cooccurrence information having intrinsic singular features, were studied and applied to personalized recommendation and document analysis for developing human-friendly intelligent information processing techniques.