支持向量机的分类算法探索与实践.doc

摘要 本文通过对支持向量机分类算法进行探索与实践，来实现对个人信用数据的分类，即个人信用评级。 支持向量机（Support Vector Machine，SVM）是在统计学习理论的基础上发展起来的一种新的机器学习方法，它基于结构风险最小化原则，能有效地解决过学习问题，具有良好的推广性能和较好的分类能力。信用评级对于银行的经营管理以及银行客户都有着十分重要的意义，个人信用评级是其中非常重要的组成部分。本文首先对支持向量机的基本思想、分类算法原理等进行比较系统的学习与探讨，再对银行个人信用数据（本文主要采用的是建行和工行的数据指标）进行分析、比较。最后对MATLAB下的Libsvm工具箱进行系统学习，实现Libsvm工具箱中SVM在个人信用评级中的应用。 基于支持向量机分类算法具有较好的分类预测能力的特性，应用MATLAB下的Libsvm工具箱，较好地实现了个人信用评级。 关键词：支持向量机；分类算法；个人信用特征数据；个人信用评级；MATLAB；Libsvm工具箱 ABSTRACT Based on the support vector machine classification algorithm to explore and practice, to achieve the classification of personal credit data, that is personal credit rating. Support Vector Machine (Support Vector Machine, SVM) is in statistical learning theory developed on the basis of a new machine learning method, which based on structural risk minimization, can effectively solve the overfitting problem, has a good Tuiguang performance and better classification ability. Credit rating for the management of banks and bank customers are of great significance, personal credit rating is one of the very important part. This paper first the basic idea of support vector machines, classification algorithms were compared between the systems theory of learning and exploration, then the bank personal credit data (This article is used in ICBC and China Construction Banks data index) were analyzed and compared. Finally, the MATLAB toolbox system Libsvm under study, to achieve Libsvm SVM toolbox in the Personal Credit Application. Based on support vector machine classification algorithm has better predictive power of the classification features, applications under the Libsvm MATLAB toolbox, to better achieve a personal credit rating. Key words: Support Vector Machine; Classification Algorithm; Personal credit data; personal credit rating; MATLAB; libsvm toolbox 目 录 1 绪论 1 2 支持向量机理论综述 3 2.1 支持向量机的相关背景 3 2.2 支持向量机的发展 4 2.3 支持向量机的相关知识 5 3 个人信用评级概述 10 3.1 个人信用评级的相关背景 10 3.2 个人信用评级的特征数据 12 4 工具箱的学习与应用 18 4.1 MATLAB的基本原理 18 4.2 Libsvm的应用 19 4