In this paper, we present a prediction method for traffics which is required for intelligent traffic management. To achieve this functionality, our method employs data mining. Data mining is used for automatically extracting meaning information from large amount of data. We prepared numerical data sets given from Metropolitan Expressway such as traffic volume, velocity, occupancy etc. Those data sets are obtained by supersonic sensors. Our method relies on two steps. The first step aligns the data sets to reduce computation costs and to understand data more easily. For instance, it uses a classification of congestions. The second step estimates prediction for traffics by using neural networks which is one of data mining method. We attempted several learning parameters and comparing has done. We had some notable results.
展开▼
