Wavelet based Feature Analysis Applied to Categorization of Electricity Data

R. Jiang, H. Tagaris, and A. Lachsz (Australia)

Keywords

--Time series analysis, wavelet transform (WT), feature extraction, modeling and simulation

Abstract

--The authors present a novel method of categorization of electricity consumers based on the energy consumption data. Wavelet multiscale properties and an optimized 3-dimensional model are used to automate and optimize feature analysis and categorization under the assumption that energy consumption data are indicators of the consumers' business category. The feature extraction scheme is carried out in both time and wavelet domains by a soft competition algorithm. A hierarchical structure in the wavelet domains is applied. Simulation results of wavelet based feature analysis and 3-D models on k-mean clustering indicate that the proposed method is effective in the categorization of electricity consumers with more accuracy, less volatility and faster convergence.

Important Links:



Go Back


IASTED
Rotating Call For Paper Image