Daily Load Forecasting Using the Self-Organizing Map

Abstract

In this paper, a daily load forecasting algorithm using the self-organizing map(SOM) method is proposed and examined. SOM is a new, powerful software tool for the visualization of high-dimensional data. It requires less training time compared to other networks such as BP learning network, and moreover, its self organizing feature can amend the distorted data. The proposed algorithm analyzes the load patterns of the past couple of years and estimates future load demand by mapping the target day using SOM. KEPCO's hourly load record obtained between 1993 and 1995 is examined to investigate the efficiency of the proposed method. It is shown that the proposed algorithm provides better forecasting results than conventional exponential smoothing method.

In this paper, a daily load forecasting algorithm using the self-organizing map(SOM) method is proposed and examined. SOM is a new, powerful software tool for the visualization of high-dimensional data. It requires less training time compared to other networks such as BP learning network, and moreover, its self organizing feature can amend the distorted data. The proposed algorithm analyzes the load patterns of the past couple of years and estimates future load demand by mapping the target day using SOM. KEPCO's hourly load record obtained between 1993 and 1995 is examined to investigate the efficiency of the proposed method. It is shown that the proposed algorithm provides better forecasting results than conventional exponential smoothing method.