An Improved ML-kNN Approach Based on Coupled Similarity

摘要：

　　ML-kNN is a well-known algorithm for multi-label classification,but it just assumes the independence of labels and instances.In fact,in the real world,labels or instances are more or less related via explicit or implicit relationships.In this paper,we propose an improved ML-kNN approach that takes the coupled similarity of attributes and labels into account,where coupling between attributes is used to find k nearest neighbors for instances and coupling between labels is used to predict the labels of unseen instances.Experimental results show that our proposed method outperforms the traditional ML-kNN.

关键词： Multi-label classification k-NN Coupled similarity

作者: Xiaodan Yang Lihua Zhou Lizhen Wang

作者单位: Department of Computer Science and Engineering,Yunnan University,Kunming 650091,China

会议类型: 国际会议

会议名称: International Asia-Pacific Web Conference(第18届国际亚太互联网大会)

会议地点: 苏州

会议语种:英文

页码: 77-89

在线出版日期: 2016-09-23（万方平台首次上网日期，不代表论文的发表时间）