一种基于多向量相似度的聚类分析方法研究

A Research on Clustering Analysis Method Based on Multi-vector Similarity

从长度、方向以及位置等方面研究了向量相似度的定义。指出了两向量间相似度衡量与多向量间相似度衡量的区别。运用AHP(层次分析法)法求出指标的权重,借鉴TOPSIS(逼近于理想解的排序方法)方法的思想,给出了一种基于多向量相似度的聚类分析方法。分别运用传统的聚类分析方法和本文给出的聚类分析方法对某工业园区的企业进行分类研究,结果表明,基于多向量相似度的聚类分析方法能减少计算量,明确分类结果的优劣,从而提高分类的科学性和有效性。

The definition of vector similarity is studied from its length,direction and position.The difference between two-vector similarity measure and multi-vector similarity measure is investigated.A clustering analysis method based on multi-vector similarity is given by using AHP method to calculate the weight of index and referring TOPSIS method.The traditional clustering analysis method and the clustering analysis method given in this paper are used to classify the enterprises in an industrial park.The results show that the clustering analysis method based on multi-vector similarity can reduce the calculation amount,indicate the advantages and disadvantages of the enterprises and improve the scientificity and effectiveness of the classification work.