基于K-means聚类算法的分析及应用

Analyse and Application Based on K-means Clustering Algorithm

聚类分析能作为一个独立的工具来获得数据分布的情况,观察每一个簇的特点,集中对特定的某些簇作进一步的分析;本文主要介绍了传统聚类算法及其局限性,然后对直接K-means算法进行分析改进,着重分析了该算法的思想体系以及它的优缺点,针对它的缺点之一提出了一种基于距离的改进策略,并将该改进策略应用到对学生成绩的分析中,实验目的是应用该算法将学生划分为合理的簇(或类)以及对聚类结果进行分析,总之实验表明了该算法的灵活性以及在此应用中的适用性.

Clustering analyse as an independent fool can get the data distribution,observe every clustering characteristic and focus on the further analyse to specifical cluster. This paper mainly introduces the traditional clustering algorithm and its disadvantages. It also analyses and improves the direct K-means algorithm. It emphasizes the algorithm＇s thinking system and its advantages and disadvantages. Overcoming one of the disadvantages, it can put forward to improve the tactic based on the distance. This tactic can be used to analyse the student scores. The experimental intention is that students are divided into rational cluster. In a word,the experiment indicate the algorithm＇s agility and validity in its application.