摘要 |
<P>PROBLEM TO BE SOLVED: To provide an electricity price predicting method capable of predicting with high accuracy. <P>SOLUTION: A price zone of electricity price is classified into a plurality of zones (S1). A fuzzy inference network (FIN) using a fuzzy self-organization map (FSOM) is constructed in a storage device by using learning data in each zone (S2). A plurality of input data is respectively input by using each constructed FIN, and attribution probability to each zone as the output of each FIN is calculated (S3). The electricity price is predicted on the basis of the determined attribution probability (S4). <P>COPYRIGHT: (C)2006,JPO&NCIPI |