| 摘要 |
<P>PROBLEM TO BE SOLVED: To achieve a classification method, for a customer load profile, which grasps the feature of the load profile storing the power consumption characteristics of an customer for electric power, and classifies customers into clusters, each of which gathers customers having an analogous power consumption pattern together. <P>SOLUTION: By the use of a customer load profile database 10 storing the load profile indicating the power consumption characteristic of a customer for electric power, a feature composed of a base pattern and each frequency component is extracted with a feature extraction processing part 11 by a wavelet transform. Depending on these features, the similarity of two load profiles is set with a similarity setting part 13. By the use of the similarity, the load profiles are classified into clusters with a clustering processing part 12. <P>COPYRIGHT: (C)2004,JPO&NCIPI |
