INCREMENTALLY FAST UPDATED SEQUENTIAL PATTERN TREES

摘要：

In the past, the FUFP-tree maintenance algorithm is proposed to efficiently handle the association rules in incremental mining. In this paper, we attempt to modify the FUFP-tree maintenance algorithm for maintaining sequential patterns based on the concept of pre-large sequences to reduce the need for rescanning original databases in incremental mining. A fast updated sequential pattern trees (FUSP trees) structure and the maintenance algorithm are proposed, which makes the tree update process become easier. It does not require rescanning original customer sequences until the accumulative amount of newly added customer sequences exceed a safety bound, which depends on database size. The proposed approach thus becomes efficiently and effectively for handling newly added customer sequences.

关键词： Data mining sequential pattern FUSP tree large sequence pre-large sequence incremental mining

作者: TZUNG-PEI HONG HSIN-YI CHEN CHUN-WEI LIN SHENG-TUN LI

作者单位: Department of Computer Science and Information Engineering National University of Kaohsiung, Kaohsiu Advanced Master Business Administration, National Cheng Kung University, Tainan, 701, Taiwan, R.O.C. Department of Computer Science and Information Engineering, National Cheng Kung University, Tainan, Institute of Information Management, National Cheng Kung University, Tainan, 701, Taiwan, R.O.C.

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 3991-3996

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）