Traffic Accident Analysis Based On K-Means

The traffic situation has become increasingly complex in an era when the society is developing at a rapid pace and most of the roads are already saturated.Problems such as road facilities,road network design,drivers' basic conditions,and behavioral norms of pedestrians are gradually exposed under this environment.Therefore,traffic accidents are a spurred by the above problems.Deep analysis of traffic accidents can help in the mining of traffic system defects,at the same time,effectively blocking some of the causes of traffic accidents can reduce the probability of accidents.Based on the road traffic system and taking the mechanism of accidents as the starting point,this paper applies the traffic accident comprehensive analysis model to road traffic accidents,and subdivides the accident comprehensive research into three directions: single factor analysis of accidents,multi-factor analysis of accidents,analysis of the accident classification.The PCA algorithm,Prior algorithm,and K-Means algorithm were selected for different levels of analysis.In order to improve the relevance of the Apriori algorithm and traffic accidents,the accidents were classified according to their severity,and at the same time,the data mining efficiency was improved through a reasonable static sampling method.Research improved the selection of initial mean points in K-Means by Combining the Canopy clustering results with the random selection method which improved the anti-noise ability and clustering effect of the algorithm.Finally,the study above analysis model was applied to the British public traffic accident data collection and analysis of traffic accidents on every level.The results show that different levels of model analysis can tap into the richer knowledge in the data,allowing more comprehensive analysis of traffic accidents and understanding of potential problems in road systems.