Automotive Product Analysis Based on MP-DP-Kmeans Clustering
Aijing Feng · 2023
Competitive product analysis is to analyze and make horizontal comparison of competitors’ products, which is an important reference basis for determining the direction of automotive product development. And clustering analysis is one of the commonly used data analysis methods, such as K-means, Density peaks clustering(DPC) and etc. However, K-means and DPC both have disadvantages, so we proposed an improved K-means, i.e., MP-DP-Kmeans. Firstly, MP-similarity is created to instead of the traditional Euclidean distance. Then, in order to make K-means obtain better clustering results, the initial clustering centers determined through DPC to solve the problem of K-means randomly selecting the initial clustering centers. Finally we apply MP-DP-Kmeans to conduct automotive product analysis.