DATA MINING FOR PREDICTING THE AMOUNT OF COFFEE PRODUCTION USING CRISP-DM METHOD

Data Mining Untuk Memprediksi Jumlah Produksi Tanaman Kopi Menggunakan Metode CRISP-DM

Abstract

The production of coffee plantations has become the leading plantation commodity with the export value of the fourth rank after oil palm, rubber and coconut. The number of coffee needs for export every year always increases, therefore it is necessary to predict the yield of coffee plants to estimate planting and anticipation that will be done so as to achieve the target. Coffee plant productivity is influenced by internal and external factors, namely the quality of the plant itself, soil, altitude and climate. The method used in this study is the CRISP-DM method and multiple linear regression algorithm to predict the amount of coffee production and determine the relationship between the variables. The steps taken are business understanding, data understanding, data preparation, modeling and evaluation. The data set that is used as many as 170 data after going through the data preparation stage produced 150 data with 5 attributes in the table. With calculations using tools, the coefficient of determination is 91.96%. That the variation in the value of the production of coffee plants is influenced by independent variables, namely the area of ​​plantations, rainfall, air pressure and solar radiation by 91.96% and 8.04% influenced by other variables not measured in this study. The results of the evaluation and validation of predictions produce good accuracy with an RMSE value of 0.3477.