Research On The Analysis Methods Of Road Traffic Accidents

Road safety is a worldwide social problem; road traffic accidents caused economic losses of about 1% to 3% of total national GDP. Car ownership in China accounts for about 3% of the world, but the road traffic mortality rate is the highest in the world over the years. To reduce traffic accidents, the protection of national property and people's lives and property, the government has to spend a lot of manpower and financial resources addressing road traffic accident-prone sections of the investigation, development of remediation measures and so on.With economic development and prosperity of motorization, the craving to high speed, high demand, high frequency, results in the process of road traffic accidents with system complexity, material performance characteristics.Therefore, it is necessary to study the intelligent analysis of the relevant theories and methods of road traffic accidents, mining accidents'regularity, hidden, causes and sudden in occurrence. The results will assist the traffic control department as decision-making. These methods enable us to facilitate the development of intelligent transportation as well as police science and technology.In this paper, I constructed a road accident information analysis system, the pattern of relationships have complied with the 3NF. It contains large amounts of information in the database, some of the fundamental factors leading to accidents, some accidents related factors, some information and identification of work-related road traffic accidents, but some of the information supporting records. Therefore, not only should we grasp the principal contradiction, but also propose accident identifying information, simplifying the complexity of the work.Rough set theory is a mathematical tools, deal with its ambiguity and uncertainty, which can be applied to reveal the relationship between inaccurate data, finding the dependence between objects and properties to evaluate the importance of the classification properties. Remove redundant data, reduct information systems to generate decision rules.Calculated by the core and reduction, rough set can be used for feature reduction, feature extraction and attribute correlation analysis. Rough set is computationally intensive, especially the reduction of the computation is a NP-complete problem. In practice, to simplify the calculation, the use of heuristic information is often the approximate reduction. It can be used as the basic set of core reduction, according to the importance of attributes one by one in descending order by adding attributes to the basic set, until the collection is a reduction to date; then checks each attribute in the collection, if deleted from the collection of a Property does not change the set of dependence on the decision attribute, this attribute will be deleted from the collection until the collection date will not change.In the calculated reduction, based on the analysis of properties using some statistics expectation, variance, by using correlation coefficient, we can choose a specific decision to any reduction, so that, to determine the direct factors associated with the accident.To facilitate analysis, we can utilize data visualization technology. Visualization and Computer Graphics, multimedia systems, human-machine interface etc. are closely related. We aim to make data, information and knowledge transfer into human perceived visual representation, from data collection to analysis of implicit and useful knowledge, so that to guide decision-makers to make scientific predictions and decision making.Visualization methods implemented in this paper including: line charts, pie charts, scatter plots, histograms, parallel coordinate method, etc. They can be directly used for analysis of road traffic accidents.Using Rough Set Theory and the related statistical indicators and visual graphics, the user can analyze the traffic accident information effectively.After analysising the accident data in the database, road traffic situation can be summarized as the following characteristics: First, serious traffic violations; Second, the National Highways are still accident-prone points; Third, weather factors; Forth, the increasing motorcycle accidents.Intelligent transportation systems facilitate the analysis for traffic accidents. With the development of Internet technology, we will enhance access to information tools and methods in the transport system greatly. Facing large-scale data, we may consider machine learning, data mining applications. Specific research will involve data preprocessing, data and relevant measure of similarity; can be used in the cluster area by clustering based on density-based clustering; in the classification, decision trees, naive Bayesian network, artificial neural Network can be used for road traffic accident data; in visualization technology, we can enhance data in three-dimensional display space.