By Topic

Toward Estimating Autonomous Neural Network-Based Electric Load Forecasters

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Vitor Hugo Ferreira ; Power Syst. Lab., Rio de Janeiro ; Alexandre P. Alves da Silva

Anticipation of load's future behavior is very important for decision making in power system operation and planning. During the last 40 years, many different load models have been proposed for short-term forecasting. After 1991, the literature on this subject has been dominated by neural network (NN) based proposals. This is mainly due to the NNs' capacity for capturing the nonlinear relationship between load and exogenous variables. However, one major risk in using neural models is the possibility of excessive training data approximation, i.e., overfitting, which usually increases the out-of-sample forecasting errors. The extent of nonlinearity provided by NN-based load forecasters, which depends on the input space representation, has been adjusted using heuristic procedures. Training early stopping based on cross validation, network pruning methods, and architecture selection based on trial and error are popular. The empirical nature of these procedures makes their application cumbersome and time consuming. This paper develops two nonparametric procedures for solving, in a coupled way, the problems of NN structure and input selection for short-term load forecasting.

Published in:

IEEE Transactions on Power Systems  (Volume:22 ,  Issue: 4 )