PENERAPAN DATA MINING UNTUK MEMPREDIKSI HARGA BAHAN PANGAN DI INDONESIA MENGGUNAKAN ALGORITMA K-NEAREST NEIGHBOR

Abstract

Food is one of the important sectors for human life, because the basic human need is food. Over time, food prices in Indonesia are often unstable. This of course has a big impact on the community and farmers. The cause of this food price volatility can be caused by several factors, such as environmental factors, pest and planthopper attacks, and dry land. The application of one of the data mining methods in the process of predicting food prices, especially predicting the price of rice in Indonesia with this prediction, people throughout Indonesia can be more vigilant in the future to prevent hunger. The benefit of predicting food prices for farmers is to prevent large losses caused by the drop in food prices. Meanwhile, the Government's goal is to maintain food security in the territory of the Indonesian Government to prevent hunger. The dataset used comes from the National PIHPS Website. The method used is the K-Nearest Neighbor (KNN) regression algorithm and for testing it uses RMSE and MAE. The results of this study, the K-Nearest Neighbor method with the regression model can predict the price of rice in January 2019 - December 2021 which is recorded every month with the addition of variable data Harvested Area (ha), Production Yield (tons) of the total experiment using range k = (2,10) then the best prediction results are by using k = 2 with MAE and RMSE for training data 52.77 and 96.40 and for testing data 55.55 and 81.64 and K = 2 parameters which have been normalized.