作者机构:
[陈宇波] College of Electronic and Optical Engineering, Nanjing University of Posts and Telecommunications, Nanjing;210023, China;[薛云; 邹滨] Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring Ministry of Education, School of Geosciences and Info-physics, Central South University, Changsha;410083, China;[周松林] School of Municipal and Surveying Engineering, Planning, Architectural Design and Research Institute, Hunan City University, Yiyang
通讯机构:
Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring Ministry of Education, School of Geosciences and Info-physics, Central South University, Changsha, China
会议名称:
International Joint Conference on Neural Networks (IJCNN)
会议时间:
JUL 06-11, 2014
会议地点:
Beijing, PEOPLES R CHINA
会议主办单位:
[Shi, Zhongwei] Guilin Univ Elect Technol, Sch Comp Sci & Engn, Guilin, Peoples R China.^[Wen, Yimin] Guilin Univ Elect Technol, Guangxi Key Lab Trusted Software, Guilin, Peoples R China.^[Xue, Yun] Hunan City Univ, Sch Municipal & Surveying Engn, Changsha, Hunan, Peoples R China.^[Cai, Guoyong] Guilin Univ Elect Technol, Guangxi Key Lab Trusted Software, Guilin, Peoples R China.
会议论文集名称:
IEEE International Joint Conference on Neural Networks (IJCNN)
关键词:
class incremental learning;concept drift;evolving data streams;multi-label classification
摘要:
Multi-label stream classification has not been fully explored for the unique properties of large data volumes, realtime, label dependencies, etc. Some methods try to take into account label dependencies, but they only focus on the existing frequent label combinations, leading to worse performance for multi-label classification. To deal with these problems, this paper proposes an algorithm which dynamically recognizes some new frequent label combinations and updates the trained classifier by class incremental learning strategy. Experimental results over both real-world and synthetic datasets demonstrate its better predictive performance.