Analysis Of Traffic Safety Based On Data Mining Technology

In recent years, a growing acceptance of car travel among peopleleads to big traffic flow and more car accidents, especially in large urbanwith dense population, traffic safety problems have become increasinglyprominent. Data Mining technology is one of the most popular methods ofdata analysis and processing in recent years. Therefore, we can study a lotof traffic accident data to analyze main factors leading to traffic accidentsbased on it, and then find out hidden valuable knowledge from largedatabase or Data Warehouse (DW). Once we obtained the main factorsleading to traffic accidents and grasped the laws of traffic accidents, theTraffic Administrative Department will be able to take more targetedmeasures to reduce traffic accidents and improve road traffic safety leveleffectively.This article will construct Multi-Dimensional Data Warehouse(M-DDW) to analyze traffic accident data based on an improved DataMining algorithm which can enhance traffic safety analysis results. The main contents of the article are:（1）By processing the raw data of trafficaccidents, extract several critical factors of the accident.（2）According tothe accident factors we already have, improve the traditionalone-dimension DW model and design the Multi-Dimensional DataWarehouses, referred to as M-DDW. The M-DDW model is able to storeand process multi-dimensional input data efficiently and obtain output withhigh satisfaction.（3）Get a Multi-Dimensional Apriori algorithm, referredto as M-DA, originated from classic Apriori algorithm based onAssociation Rules of Data Mining. Classical Apriori algorithm can onlyhandle input with low dimension, we have to improve the originalalgorithm to adapt to the new warehouse model when considering themulti-dimension nature of traffic accident factors.（4） Create anexperimental platform to construct M-DDW model, then input the trafficaccident data to finish independent analysis based on the M-DA algorithmto derive the degree of relevance of between all the critical accidentalfactors and accidents and get the valuable association rules. Theseinformation will help our traffic administrator to carry on efficient trafficsafety analysis.