Short term load forecasting using correction bilinear neural network (part ii)

Abstract

This paper presents a new approach of training data structure of neural network. Neural network is bilinear neural network [1]. Results show that the proposed forecasting method could provide a considerable improvement of forecasting accuracy, reducing neural network forecast error, especially if having factors influencing: daily factor, weather factor… The practicability of the proposed approach has been confirmed through a comprehensive study based on the data provided by HCMC Electricity Company.