| With the development of economy and the improvement of people's living standard, taxies take a large proportion of city passenger transportation. Meantime, a lot of traffic accidents come. Therefore, the taxi management must be reinforced and the happening of the traffic accidents must be lesson. Database has been fully used in traffic management area. So a large number of data are accumulated in this area. What's more, a lot of data which are valuable and have potential association exist in the database. It has become an important scientific research task to apply data mining technology to transportation area and use the technology to mine the traffic accidents data and has received much attention from both domestic country and abroad. The main part of the thesis is to research the clustering analysis and association rule of the data mining technology and to apply them to traffic accidents analysis system.Firstly, through combining the present situation and its developing trend of both domestic and abroad, the data mining technology and its algorithm are discussed systematically. According to the characteristic of taxi traffic accidents, data mining model is constructed. Then the demand analysis of the system and system design are preserted in detail. Emphasis is laid on methods used to process the accident data on the data preprocessing stage.The core of the thesis is how to apply data mining technology to taxi traffic accident analysis system. K-means algorithm for clustering analysis and Apriori algorithm for association rule are deeply studied. The innovative part of the thesis is that the problems existing in algorithm are analyzed and the original algorithm is improved. The experiment prove that the improved algorithm is better than the original one. The association rule and the improved on clustering analysis algorithm were used to mine the traffic accident data and the results are analyzed. The validity of the system is proved and the purpose of the experiment is reached. |
PDF Full Text Request