Research And Design Of Traffic State Classification

Traffic relates with our daily live but traffic congestion has become a problem many large and medium-sized cities facing.How to ease the congestion and optimize traffic has become an important problem we have to solve. In terms of hardware, we can improve the capacity of the vehicle by speeding up and improving the road infrastructure. In terms of software, we can use computer and data processing technology to search useful information from massive traffic data, such as traffic law, to improve the level of traffic operation and management.We can find previously hidden, contained, useful information or knowledge form the data by data mining technology. The knowledge always can be used in many fields, such as economy trend analysis, production targets set, and market positioning, to guide and help make the decision. This paper uses data mining to process traffic data and obtains a series of analysis result, having some help on easing urban traffic congestion and optimizing road.We proposed a method of short-term traffic state classification in the paper. The method is for5minutes intervals and based on the statistical characteristics of the traffic parameters. There are three main steps of the method:firstly, classify the parameters separately by the clustering algorithm according to their statistical characteristics; secondly, divide traffic state by the fuzzy combination of the parameters and build traffic state classification model; thirdly, partition the traffic state through the model. And through some examples, the algorithm has high effectiveness.