基于K-means的驾驶行为离散化特征聚类分析与研究OA

To explore potential feature relationships in continuous driving behavior data,this paper collects continuous driving behavior data of actual transportation vehicles.Firstly,through corresponding preprocessing and feature extraction,obtain continuous driving behavior data of the corresponding vehicle during the corresponding time period;secondly,the K-means clustering method,which has stable performance in both discrete standard datasets and continuous noisy datasets,is used to make discretization clustering processing and analysis on driving behavior data;finally,three representative driving behaviors are obtained:"Steady Driving""Impulsive Driving""Dangerous Driving".In addition,analyzing and studying the various hidden features in driving behavior provides a strong basis for further correlation analysis in data mining based on driving behavior data.