基于k-近邻法地分类器地研究及实现.doc

基于K-近邻法的分类器的研究与实现 摘要 模式识别的目的就是对未知的样本，判断它所在的类别。人类的模式识别能力使得人们可以很好的认识周围的环境并与之交流，如果计算机也具有类似的能力，那么其智能程度将会大大提高，可以发挥更大的功能，更好的为人类服务。本文的研究课题就属于计算机模式识别领域。 分类器是模式识别系统的重要组成部分；也是机器学习的重要研究领域。本文主要研究对象是KNN分类方法，运用K近邻法(K Nearest Neighbor)对数据进行分类，并对分类结果进行比较研究。 本文的研究工作主要探讨基于K-近邻法的分类器的实现，主要集中在K-近邻法的理论分析，算法实现。本文首先介绍了数据挖掘的目的、意义及现状，阐述了K-近邻算法在数据挖掘中的地位和作用，然后对K-近邻法进行了详细的研究与分析，并且实现基于K-近邻法的分类器。 本设计采用SQL Server 数据库系统和c#.net开发工具进行分析研究。 关键词：模式识别；数据挖掘；机器学习； K-近邻法；分类器  THE RESEARCH ACHIEVE OF CLASSIFIER BASED ON THE K-NEAREST NEIGHBOR ALGORITHM ABSTRACT The purpose of pattern recognition is judge it in the category for the unknown sample. The pattern recognition capabilities of human can make it a good understanding of the environment around and exchange with them, If the computer also has a similar capability, its smart level will greatly improve ,the level they can play a greater role and better service to humanity. This research on the subject is a kind of computer pattern recognition. Classifier is an important component part in pattern recognition system;it is also an important research in the area of machine learning.This paper mainly targets KNN classification methods, using k-nearest neighbor for data classification, and compared the results. This article research on the achieve of classifier based on the k-nearest neighbor algorithm.Mainly concentrated in the k-nearest-neighbor theoretical analysis and algorithm .First of all,I introduce the purpose、meaning and recent development of data mining.and expatiate the status and function of k- nearest neighbour in this field.then research and analysis to the k-nearest-neighbor detailed and achieve the classifier based on k-nearest-neighbor. I design this program with SQL Server database system and c #. net development tools for analysis and study. Key words: pattern recognition; data mining, machine learning; k nearest neighbour; classifier  目 录 1 绪论………………………..………………………………….………………………….1 1.1 课题背景及目的………………………………………………………………………1