Food Commodity Price Prediction in East Java Using Extreme Learning Machine (ELM) Method

Abstract

Fluctuations in food commodities price in East Java Province cause various negative impacts when there are significant changes. To avoid this problem, it is necessary to predict food commodities prices to prevent high price increases. This study aims to apply the Extreme Learning Machine (ELM) method to predict the price of staple food commodities in East Java Province and measure the performance of the ELM in predicting staple food commodities price. The ELM is a method develop from feedforward Artificial Neural Networks (ANN) with one hidden layer or commonly called Single Hidden Layer Feedforward Neural Networks (SLFNs). The prediction process of staple food commodities is carried out using 3 data features, 7 neurons, and composition of training and testing data is 80%: 20%. The results showed that the average level of prediction accuracy for all staple food commodities was 98.79%. This shows that the prediction error is very low, ie the predicted results approach the actual value.