PFCM聚类算法的可视化实现与应用研究

发布时间：2019-02-12 09:09

【摘要】：当代社会的信息化程度高度发达，许多传统行业也已加入了数字化与网络化的大潮之中，如何整理并利用数据，如何从数据之海中挖掘提炼出具有商业价值的信息及规律，已经成为亟待解决的问题。由此，数据挖掘成为了信息技术中一个非常重要的研究领域。作为数据挖掘的一种重要方法，聚类分析被广泛应用于各种领域之中。聚类分析是指用数学的方法研究和处理对象，并给出对象的合理分类。其中，C-均值聚类算法是一类应用非常广泛的聚类算法。本文主要研究C-均值聚类算法的可视化实现，包括硬C-均值聚类、模糊C-均值聚类、可能性C-均值算法以及可能性模糊C-均值聚类算法,并研究如何将这几种C-均值聚类算法应用到实际的房地产市场分析之中。本文介绍了研究的背景及国内外的发展现状，阐述了本研究的目的与意义；给出了模糊集合及其运算原理，为引入模糊聚类以及可能性模糊C-均值聚类算法做好准备。本文详细探讨了四种C-均值聚类算法的算法原理及核心步骤，并使用实验数据集对上述算法进行验证，分析探讨了各种算法的聚类性能及聚类准确度，针对每种算法的适用范围及优缺点进行了总结。最后，本文给出了这几种C-均值聚类算法在我国35个大中型城市住宅市场分析中的应用实例，，针对房地产数据的聚类结果给出了具有实际指导意义的分析与结论。
[Abstract]:The information level of contemporary society is highly developed. Many traditional industries have joined the tide of digitization and networking. How to organize and use the data, how to mine and extract the information and rules of commercial value from the sea of data, It has become an urgent problem. Therefore, data mining has become a very important research field in information technology. As an important method of data mining, clustering analysis is widely used in various fields. Clustering analysis refers to the study and processing of objects by mathematical methods, and the reasonable classification of objects. Among them, C-means clustering algorithm is a very widely used clustering algorithm. This paper mainly studies the visualization realization of Cmean clustering algorithm, including hard Cmean clustering, fuzzy Cmean clustering, possibility Cmean algorithm and possibility fuzzy Cmean clustering algorithm. And how to apply these C-means clustering algorithms to the real estate market analysis. In this paper, the background of the research and the development situation at home and abroad are introduced, the purpose and significance of this study are expounded, and the fuzzy set and its operation principle are given, which is ready for the introduction of fuzzy clustering and the possibility of fuzzy C-means clustering algorithm. In this paper, the principle and core steps of four C-means clustering algorithms are discussed in detail, and the experimental data sets are used to verify the above algorithms. The clustering performance and accuracy of these algorithms are analyzed and discussed. The application scope, advantages and disadvantages of each algorithm are summarized. Finally, this paper gives some examples of the application of the C- mean clustering algorithm in the residential market analysis of 35 large and medium-sized cities in China, and gives a practical analysis and conclusion on the clustering results of the real estate data.
【学位授予单位】：中国地质大学（北京）
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP311.13

【参考文献】