Penerapan Metode Euclidean Probality dan Confusion Matrix dalam Diagnosa Penyakit Koi

Abstract

Koi fish aquaculture is still susceptible to fish’s diseases. The farmers’ limited knowledge about koi fish diseases, especially the novice farmers, to overcome this problem which has impact on the koi fish harvest to be decreased, and even tends to losses for the farmers. The creation of an AI system for diagnosing koi fish diseases using “the euclidean probability and confusion matrix” method is needed as a solutions for farmers to discover the solutions if their fish are get diseased. AI system with “euclidean probability and confusion matrix” methods created using 10 diseases and 23 symptoms data. Based on that, 10 rules were obtained. Diagnosis of koi fish disease is carried out by entering the symptoms of koi fish disease and calculating the percentage based to the appropriate rules. After the disease was picked based on largest percentage. After obtaining a diagnosis of the disease, testing using confusion matrix is performed. The test results from 93 test data show a precision level of 93.67%, a recall rate of 88.21%, an accuracy level of 95.91%, an error rate of 4.09% and an f-measure level of 85.27%.