Short-term load forecasting based on rough set and principal component analysis

The attribute reduction methods in neural network forecasting are analyzed.In order to reduce the inputvariables and minimize the network structure,an integrated approach is suggested based on rough set theory andprincipal component analysis.The load data of some area is analyzed.The corresponding forecast models areestablished.The influential factors are simplified.The BPnetwork isemployed for simulation test.Results indicatethat this method is effective.