Artificial Neuron Network-Based Prediction of Fuel Consumption in Antananarivo

Abstract

The ability to modeling fuel consumption forecast is important to improve fuel quantity suitable for users (individuals, businesses, etc.) to prevent fuel shortages. Kerosene, super fuel and gas oil consumptions forecasting models for Antananarivo region were developed. Model is based on Artificial Neural Network (ANN) learning which used advanced machine learning techniques using backpropagation algorithm. Successful time series and trend patterns given by the three ANN prediction models were presented. To predict 18 months fuel consumption in Antananarivo, ANN models’ accuracy reaches more than 95% accuracy.