Artificial neural network method for forecasting price of purebred chicken egg in East Java

Purebred chicken eggs are a staple food source commodity that is widely chosen by the community because it has an affordable price with high protein content and sufficient availability. However, even though availability of purebred chicken eggs is sufficient, there is still a problem, namely very fluctuating prices. Efforts to read these uncertain market conditions are urgently needed for decision-making by producers, consumers, and even the government, one of which is by forecasting prices. Forecasting can be used to predict future conditions by looking at trends in past conditions. This study uses Artificial Neural Network method with a feed-forward backpropagation algorithm. The purpose of this study is to determine price forecasting architecture using ANN and forecast price of purebred chicken eggs in East Java in 2024 and 2025. The research data used is monthly price of purebred chicken eggs in East Java province from 2018 to 2024. The results of study produced best architecture, 12-12-1. The resulting training Mean Square Error value was 0.00099947 with an accuracy of 98.01%, MSE of the test was 0.023797 and MAPE value was 0.164254. Based on MAPE values generated by forecasting using Artificial Neural Network, it has a good level of forecasting ability.