Spatial Neural Network for Forecasting Energy Consumption of Palembang Area
Muhammad Rifan, Daryanto, Agung
Universitas Negeri Jakarta
Abstract
Spatial Artificial Neural Network (ANN) is proposed as a new approach to determine the level of energy consumption in the future in the Palembang region, South Sumatra. Back-propagation is used to train Artificial Neural Networks. Population size, Gross Regional Domestic Product (GRDP), economic growth, Household Energy Consumption spatially per sub-district is used for this estimate. Data for 2012-2016 is used to train Spatial Neural Networks (ANN) and Data for 2017-2018 is used to validate this new approach to electricity demand prediction. The proposed Spatial ANN approach provides relatively good predictions about energy demand in acceptable errors and high accuracy for electricity demand predictions.
Keywords: energy forecasting, spatial, neural network
Topic: Electrical Engineering