Recently, short-time traffic flow forecasting becomes a research hotspot, because only with realtime and accurate predict of traffic flow, can assure traffic control, event detection and traffic-routing. For an intersection, there exists some algorithms, but each has its special applied situation. So a new kind of algorithm with "no parameter", portability and high accuracy is in need. For a traffic network, So far many algorithms have been proposed, but problems in accuracy and timeliness still remain to be the major obstacle for their successful applications. For example, presumed human travel habit and vehicle turning probabilities at intersections have greatly limited the use of dynamic assignment algorithm. Based on data mining tech, this paper following researches: 1) Traffic data screening algorithm based on data mining Considering valve value and traffic flow theory, this paper proposed a kind of screening method based on space and time verification. Considering effect of traffic event, this paper proposed a kind of estimation for lost data based on time series, 2) Single point traffic flow forecasting based on data mining First three improvements are put forward which are effective traffic state vector selection method based on self-association analysis and association analysis, improved variable K search method based on "dense degree" and advanced data structures based on dynamic cluster method and hash-function transformation. And then based on this algorithm and analysis of a whole day traffic flow curve, a comprehensive method is proposed. 3) Traffic network flow forecasting based on data mining After analysis of traffic flow of each link in a traffic network with data mining tech, relationship among links and association rules among links are educed. Then with these rules and relationship, a new network flow forecasting method is proposed. 4) Traffic supportive routing algorithm based on CBR In order to improve the timeliness of the current traffic routing algorithm, Case Based Reasoning(CBR) is applied to support traffic routing. Through searching similar history event to the current event, realtime routing answer can easily got.
修改评论