Abstract:

Aiming at the fuzziness of supplies demand in the process of low altitude rescue, a CBR(case based reasoning) prediction algorithm based on K-means is proposed. Firstly, the attributes of case are reduced through rough set,the weight values of attributes are calculated. Next, a cluster analysis is made on simplified case through DB Index K-means algorithm; then, correlation coefficient is calculated between the current case and each case in the nearest group, retrievaling the target case with maximum similarity. Finally, according to the supplies of target case, demand of the current case is obtained. A real seismic data is conducted to compare the accuracy of the current approach and genetic optimization BP algorithm. Result shows that the K-means CBR algorithm has higher precision.

Key words: K-means, CBR, rough set, demand forecasting