An experimental research of accident report using data mining

Abstract

Traffic accident rates are increasing extremely every year. Though engineers and scholars use the latest technologies to build safe vehicles, accident reports are inevitable. Even though the vehicles are designed with state of the art technology, the accidents are unavoidable. The necessity of developing a model for predicting the dangerous crash patterns that classifies automatically the type of severity of the injuries by using a study of traffic accidents are immense. These roadway models and behaviours are cooperative in building traffic control measurements. To get the highest possible decrease of the accidents with limited economic resources, there should be a detailed study about unbiased and technical surveys and also the root cause of the accident is vital. Data mining is a technique in which from the enormous amount of data in database, hidden patterns will be taken out. It is mainly applied in detection and prediction, surveillance, and fraud analytics, etc. We can learn and recognize about complex patterns by using machine learning techniques in data mining. We can make intelligent decisions and predictions using the customized and available data. In this paper, accident exploration and traffic analysis can be proposed by certain data mining techniques. The idea is to collect traffic accident dataset and to apply data mining techniques. The type of damages are classified.