数据挖掘中信息颗粒及其构造

为在海量的数据仓库中搜寻稳定的、有意义的且易于表示的知识模式 ,简要介绍了数据挖掘的任务、分类及智能计算方法的作用 ,提出了信息颗粒的概念 .信息颗粒的引入 ,使人们更容易地理解各种复杂数据所隐含的意义 ,对数据进行模块化处理 ,从而降低运算的复杂性 ,起到信息综合、信息压缩的作用 .另外 ,根据模糊性理论构造和管理信息颗粒 ,用模糊集对信息颗粒进行量化表示 ,计算信息颗粒度 ,针对具体问题根据兴趣度灵活而有效地选择信息粒度 ,使它们有效地覆盖所研究的整个数据领域 .最后 ,讨论了颗粒间的各种关系 ,并给出了信息颗粒在数据挖掘中的进一步作用 .

Information Granule in Data Mining (DM) and Its Construction

Data mining is searching for stable, meaningful, and expressible knowledge patterns in a huge data bank. The authors state the main tasks and classification of DM and the intelligent computation methods in DM, and propose the concept of information granule. The intruduction of information granule can make the implied sence of the complex data understood easily, can make them modularized, and can reduce the complexity of treating them. So the data information is synthesed and compressed. The construction and management of the information granules are based on fuzzy theory, and the quantitative expression of them is done by means of fuzzy sets. Granularity can be diversified in terms of the interest degree of a problem in order that the information granules can cover over all the studied data. Finally, the various relationship among the granules and the further effects of the information granules in DM are discussed.