Implementation of Backpropagation Artificial Neural Network for Food Price Prediction in Majene Central Market

Food has a fairly high price and the stability of food prices can affect entrepreneurs and the community in meeting their daily needs. This is often seen as a sudden increase in prices (extreme). Therefore, it is necessary to have precise and accurate forecasts or predictions to assist local governments in taking the initial steps in efforts to stabilize food prices. Artificial Neural Networks (ANN) can be used to predict future food prices using the Backpropagation Algorithm. Sental Market is one of the trading centers for daily necessities in Majene district, West Sulawesi. The study used data taken from the Office of Cooperatives, SMEs, Industrial Trade, Kab. Majene, in the form of food price data per week. The research aims to assist the Majene Regional Government (PEMDA) in taking initial steps / policies to stabilize food prices. The system is designed to predict food prices by applying a Backpropagation Neural Network, then reviewing the accuracy obtained in the food price prediction system for each commodity. The results of the study used a Backpropagation Neural Network pattern with a total data of ±156 for each commodity. The results of the study used N.Input of 2, N.Hidden of 3, and N.Output of 1. While the parameters used were Alpha of 0.3, an error tolerance of 0.001, and a maximum iteration of 100. The highest accuracy in the prediction of Commodity Rice Prices was 98.47 with the computation time for the training and testing process being 1.69 and 0.004 respectively