自组织数据挖掘方法研究综述  

Review on improvements of self-organizing data mining

作　　者：张宾[1] 贺昌政[1] 

机构地区：[1]四川大学管理科学系,成都610064

出　　处：《哈尔滨工业大学学报》2006年第10期1719-1723,1727,共6页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金资助项目(70271073)

摘　　要：从对自组织数据挖掘方法(SODM)本身的改进和SODM与其他建模方法结合两方面介绍了SODM的研究进展.对SODM本身改进的方面包括模型表示、选择准则和搜索方法.分析SODM与神经网络、遗传算法、模糊法则归纳法以及模拟方法的结合.在此基础上,指出SODM改进方向应该在改善算法本身的基础上,将其从单纯的数据挖掘技术扩展为知识发现的全过程.Advances of improvements on self-organizing data mining （SODM） are reviewed from amelioration on itself and its combination with other modeling techniques. Amelioration on SODM includes study on its model representation, model evaluation criteria and search method. SODM＇ s integration with other modeling techniques is analyzed, involving with Neural Network, evolution algorithm, Fuzzy Induction Rules and simulation. It is pointed out the direction of SODM＇ s further development will be to broaden it from mere data mining method to effective Knowledge Discovery tool. Some suggested future work is listed.

关 键 词：自组织数据挖掘方法 数据分组处理方法 数据挖掘 知识发现 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]