Research of the Comprehensive Forecast of Hailstone

Rough Set Theory was used for data mining based on the characteristic database and can form a knowledge database for hailstone recognition to establish a single model for hailstone forecast. Thus the comprehensive hailstone forecasting model was formed. Firstly the rules discovered from Apriori algorithm were used to eliminate the interference, Secondly the integrated knowledge database was formed by the combination of the rules obtained from the Rough Set Theory and the FP-tree algorithm for the comprehensive forecast. Certainty factor model was adopted to solve the rule conflict which occurred as the number of the rule increased. The experimental results show that the comprehensive forecast model improved the accuracy of hailstone recognition and good results were achieved in practical operation.