A combined collective classification method for sparse data
Journal of Information and Computational Science(2011)
摘要
Collective classification in networked data has become an important and active research topic, it has a wide variety of real world applications, such as hyperlinked document classification, protein interaction and gene expression data classification, recommending system analysis. However, previous work has shown that the performance of collective classification can degrade when there are too few labels available. In this paper, we firstly present a novel mix collective classification framework that combines network structure features with label distribution of nodes to perform this task, then propose a new method to implement this framework. Through experiments on four real world datasets, we demonstrate that our method outperforms the other state-of-art approaches for this problem. © 2009 by Binary Information Press.
更多查看译文
关键词
Collective classification,Link pattern,Networked data
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络