Data Mining Of The Road Traffic Accident Based On Association Rules

Our country is the highest country of death toll on the road trafficaccident, having been occupying the first place in the world all the time forseveral years in succession. In order to reduce the road traffic accident andensure national wealth and people's safety of life and property, our countrycost a large number of human and financial resources to check black spot ofthe road and make countermeasures every year. The conception of theIntellectual Traffic System is to guarantee the security of the road traffic and tomake the road transportation reach the optimum disposition with the advancedinformation technology and communication technology. What the intellectualtraffic is studied is mainly studying the characteristic of each regional trafficaccident of the road, the factors of people or vehicles or roads that cause thetraffic accidents, the countermeasures that can improve the safety of roadtraffic, etc.. With the development of market economy, a lot of theoreticalquestions and practical problems need to be solved in aspects of the regionaltraffic control, the urban traffic control and management, so the intellectualtraffic is studied has very important meanings.The technology that this text studies belongs to the Intellectual TrafficSystem. It utilizes the technology of association rules to find the potential andvaluable laws from the database which records the data of traffic accidents,which is used to guide the administrative decisions of the traffic, to stoppotential accidents and to reduce accidents to take place. This text makes useof the improved algorithm of multidimensional Apriori to realize the schemeof decision and analysis.Data Mining is a course of excavating meaningful knowledge mode fromthe noise and incomplete database. The proposition of data mining methodsmake people have abilities to realize the sterling worth of the data. Theeffective data mining from large-scale database has put forward a largenumber of demands and enormous challenges to researchers and developers.Association rules is the main data mining methods, which reflects theinterdependence and relation among one thing and other things. It means therules that the support and confidence accord with the given value in data sets.The classical algorithm of association rules is Apriori, which is name by thepriori knowledge about the characteristic of frequent itemsets. A fundamentalproperty that contains in the algorithm of Apriori is that any subsets of afrequent itemsets should be frequent. In virtue of certain professionalknowledge, association rules can be directly used to analyse the datacausalities, to study further and to make predictions of rules. Finding relatedrelations from a large number of data is extremely useful in such fields as themarket orientation, the decision analysis and the commercial management,etc..Chapter 3. This text adopts the technology of data mining in artificialintelligence to analyse relevant data, builds data model based on the theory ofassociation rules and links with the corresponding Geographical InformationSystem of traffic, which form a intelligent and visual software system. Thesystem contains four pieces of module, which are coordinate input, black spotinvestigation, statistic analysis and decision analysis of traffic accidents . Thedecision and analysis module is the focal and difficult point in the project, it ismain research work of this text too. The decision and analysis module is facingpolicymakers in traffic supervision department, which utilizes the theory ofassociation rules to analyse the past data of traffic accidents and finds outrelevant potential factors that can cause traffic accidents. That can offer certainbasis for policymaker make decisions.Chapter 4. There are its specific and accidental reasons, and subjected tothe domination of various kinds of other factors in the emergence of the trafficaccidents. If we can find their inherent laws of the connection from theinducements of a large number of traffic accidents, control and interveneartificially, make the cause factors lack, it can reduce the emergenceprobability of the traffic accidents. This text is proposed a new method toanalyse the data inner link of the traffic accident. It disposes the trafficaccident data, distills and analyses the traffic accident laws by associationrules in data mining.In allusion to the fact of the traffic accident, this text improves theclassical single-dimensional and single-layer algorithm of Apriori and utilizesthe new multi-dimensional and mutli-type algorithm of Apriori based onassociation rules. It organizes a large number of complicated and disorderedtraffic accident data to be information, analyses various kinds of complicatedrelevant factors. Here provides the definition of property of the road trafficaccidents (PRTA) and adopts star-type the full -connection data model to build.The first algorithm step is to find out frequent itemsets that accords with theminimum support. The improvement of the algorithm is through improving theproduction way in candidate itemsets, which is named the function ofgen_candidate(). The second step is to utilize frequent itemsets that producedabove to get expected rules. At the same time, It utilizes the important nature,which is "The support of any data itemsets is always less than or equal that ofsubsets."That can reduce the range of search, so it improves the efficiency ofthe algorithm.Chapter 5. This text adopts the research approach that consists with thetheoretical research, the algorithm experiment and the practical applicationclosely. According to the Ministry of Public Security's standard database,system disposes the systematic environment to develop, and utilizes improvedmulti-dimensional Apriori algorithm. System utilizes the traffic accident dataabout the urban area (the traffic police group of Nanguan, Kuancheng,Chaoyan, Erdao, Qichechang, Luyuan) from Dec.21.1998 to Jul. 20.2004,which provided by the Center of Traffic Control in Changchun, to test. Itdraws and analyses association rules among the road traffic accident data inthe urban area of Changchun, which produces a large amount of strongassociation rules that accord with support and confidence to analyse the mainreasons that cause the traffic accident effectively. So that can propose thescheme of significant management and precautionary measures for thepolicymaker.Problems that can be further investigated in this text:The method that is used to draw and analyse association rules of thetraffic accident data in the urban area of Changchun in this text, whichrepresents the use value of the drawing technology of association rules. Wecan build the data model of property of the road traffic accident in more detailso that the model can contain more information and can carry on more detaildata association rules to analysis. Meanwhile , we can adopt reasonable way tostore the result of the black spot investigation in order to use it in the moduleof the decision and analysis directly to analyse, So the arrange way in the forth...