Input Features Selection using RReliefF Algorithm for Electricity Demand Forecasting

Electricity Demand Forecasting (EDF) allows the power utilities to organize their resources in advance and to execute demand response strategies more effectively, which consequently enables several features such as better uninterruptible service to customers with affordable electricity tariffs and higher sustainability. This paper deliberates the implementation of RReliefF algorithm for selection of input features required for EDF. The proposed day-ahead EDF models are developed by sequence-to-sequence regression using deep learning technique with parallel input and output EDF framework. The real time data is obtained from Tata Power Delhi Distribution Limited (TPDDL)which has been utilized for this research work.