The Clustering Analysis on Public Health Data with Missing Values Based on Dimension Reduction Methods
Honghao Zhao, Weiyi Ding, Fang Ye, Weimeng Yuan, Hangyu Chen · Atlantis Highlights in Computer Sciences/Atlantis highlights in computer sciences · 2023
With the development of medical information digitization, machine learning techniques have become a popular method of mining medical health data for hidden information and knowledge.Health data from normal medical checking is usually limited.However, public health data from magnetic resonance (MR) are usually high-dimensional data with missing values.This paper presents a clustering analysis of such health data with a series of steps, including filling missing values, dimension reduction, and clustering to provide a framework with potential solutions to the problems of missing value and high data dimension.Our results show that the UMAP method is the most effective one for dimension reduction, and the K-means clustering method works well in most cases.