基于I-Miner及S语言的分类算法研究

Classification Algorithm Research Based on I-Miner and S Language

分类是一种重要的数据挖掘技术,其目的是根据数据集的特点构造一个分类函数或分类模型(也常常称作分类器),该模型能把未知类别的样本映射到给定类别中的某一个。通过介绍I-Miner下的数据挖掘实验方法,并利用S语言做成的脚本,实现了在I-Miner中没有实现的算法,主要介绍S语言实现分类算法中的K-最邻近算法,通过对不同数据集的实验,验证了K-最邻近算法的特性,并以此为今后改进算法做好基础。

Classification is one of the most important technology in data mining and aims to create a classification function or model（also called categorizer）, which can classify an unknown sample into a well-defined class via characteristic of data sets. This paper introduces a new method of data mining based on I-Miner and uses a script made by S to describe an algorithm that never implemented in I-Miner. The K-Nearest neighbor classifier algorithm made by S is mainly introduced and its characteristic is proved by testing on different data sets. And consequently we lay the foundation for improving algorithm for the future.