Preprocessing data for short-term load forecasting with a general regression neural network and a moving average filter

This paper proposes a filter based on a general regression neural network and a moving average filter, for preprocessing half-hourly load data for short-term multinodal load forecasting, discussed in another paper. Tests made with half-hourly load data from nine New Zealand electrical substations demonstrate that this filter is able to handle noise, missing data and abnormal data. © 2011 IEEE.

Keywords

Artificial Neural Networks, Moving Average Filter, Short Term Load Forecasting, Signal Processing, Training Dataset, Abnormal data, Electrical substations, Filter-based, General regression neural network, Load data, Load forecasting, Missing data, Moving average filter, New zealand, Forecasting, Neural networks, Signal processing, Sustainable development, Electric load forecasting