Adaptive Neuro-Fuzzy Inference System implementation for farmer's term of trade forecasting in West Sumatra

Abstract

The farmer's term of trade (NTP) is used by the Central Bureau of Statistics (BPS) as one of the indicators for measuring the level of welfare or purchasing power of farmers. To prepare preventive measures when the NTP index falls from the previous period, the relevant parties need to predict NTP for the coming period. The purpose of this study is to measure the performance of Adaptive Neuro-Fuzzy Inference System (ANFIS) in predicting NTP of West Sumatra Province in the coming month. The data used are monthly NTP data of West Sumatra Province obtained from the BPS website www.sumbar.bps.go.id from 2013 to 2016 for the network training process. Model evaluation was done by comparing the test results with actual data in 2017. Forecasting systems are divided into two types, namely time series model and the multivariate model. The test results show that the time series model has the smallest RMSE of 0.3430 for the training set and the RMSE value is 1.4570 for the testing set.