Sistem Pakar Penentuan Jenis Budidaya Ikan Air Tawar Berdasarkan Lokasi dan Kualitas Air

Abstract

This research developed an expert system designed to record and use knowledge, experience, and expertise of the experts in aquaculture to identify what types of freshwater fish farming based on location and existing water quality. K-fold cross validation method is used for sharing subset of data train and data test with three replications for each training and testing. Nominal distance is used to calculate distances on the nominal parameters, while Euclidean distance is used to calculate numerical parameters that will be combined using aggregate formula. The resulting distances are combined by aggregate formula. The classification method used for determining the appropriate class type is K-Nearest Neighbor that give the best average of accuracy at k = 3 with normalization, with a value of 97.41%.