| Hailstone is a main disaster in the Meteorological area, which threats to life and property. Because the hail echoes arises suddenly, moves fast and dissipates very quickly, it is difficult to forecast. So the research on hailstone echo automatic forecast system can effectively reduce the damage of the hailstone.This paper introduced the approach to knowledge discovery in time series database (TSDB), which based on Mutual Information of Shannon's Information Theory and the radar echoes TSDB. The results of knowledge discovery in TSDB can be used to predict time series behavior in the future, and was presented to the forecast of hail echo.The main work of the paper included:(1) The radar echoes TSDB was constructed. According to digital image processing techniques, the features of color, texture and shape were extracted and calculated from the radar image, and then the features were stored in the generate database. On discussing and analyzing the different construction of Special Database, time-series database from general database was remodeled, which provides the data resource to the research of the hail echo forecast.(2) The construction of MINM was detailed, which is used to discover knowledge. Using the concept of Information Theory and Mutual Information, calculated the difference between priori and post entropy to detect the associating degree between input attributes and target attributes, which controls the network configuration. The network can be used to deduct associating rules of process of hailstone.(3) With the optimization of rules which were obtained from data mining, an approach of automatic hail echo forecast was proposed and implemented. Theoretical analysis and experiment showed that this method was able to forecast the hail echo on the radar image effectively.In conclusion, the research of this paper can be applied to the practical situations, and has made great contribution to the hailstone automatic forecast system. |
PDF Full Text Request