期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：19
页码：5081
出版社：Journal of Theoretical and Applied
摘要：The real-world application has grown in need of heterogeneous data classification for almost all kind of datasets. The complexity in learning a class for a single object which is associated with multiple label sets is a key problem for multi-label datasets. Existing methods might be unfavourable for classification as each label consists of specific features characterization. This paper propose a One-To-k (OT-k) Label learning method through exploiting the labels characterization and using association rules to discover label dependencies for the classification. The main objective is to find One-Label which will be highly suitable for class suggestion using a OL-Prediction Table and k-labels to constructs a patterns of labels to deal with the multi-label database classification. The efficiency of OT-k is verified against other multi-label learning algorithms. The result analysis shows an improvisation in different case studies being performed.
关键词：Label Learning; One-To-K; Pattern; Classification; Association Rules; Multi-Label Datasets
