Load Time-Series Classification Based on Pattern Recognition Methods

Book Description

In this chapter pattern recognition methodologies for the study of the load time series were presented. Specifically, the first methodology deals with the classification of the daily chronological load curves of each large electricity customer, in order to estimate his typical days and his respective representative daily load profiles. It is based on classical pattern recognition methods, such as k-means, hierarchical agglomerative clustering, Kohonen adaptive vector quantization, mono-dimensional and bi-dimensional self-organized maps and fuzzy k-means. The parameters of each clustering method are properly selected by an optimization process, which is separately applied for each one of six adequacy measures. The latter are the mean square error, the mean index adequacy, the clustering dispersion indicator, the similarity matrix, the Davies-Bouldin indicator and the ratio of within cluster