Implementasi Algoritma K-Nearest Neighbor (K-NN) dan Single Layer Perceptron (SLP) Dalam Prediksi Penyakit Sirosis Biliari Primer

Abstract

Primary biliary cirrhosis is a chronic cholestatic liver disease that can lead to liver failure. The majority of individuals who suffer from this disease are women. Primary biliary cirrhosis is recorded as contributing to mortality worldwide with a percentage of 0.6% to 2.0%. However, so far, randomized trials have shown that some immunosuppressant or immunosuppressive drugs do not play a major role in patients with primary biliary cirrhosis. Therefore, early detection is important to start treatment and planning for appropriate medical needs. The results of the processing accuracy with the K-NN algorithm of 76.2% and the SLP algorithm of 63% using the Percentage Split method show that the K-NN algorithm is better for early detection of primary biliary cirrhosis. The K-Nearest Neighbor algorithm is able to perform early detection of primary biliary cirrhosis with a precision of 77% and recall of 75% with the hope that the percentage of mortality worldwide can decrease. However, the K-NN algorithm is not superior in retrieving information in patients with primary biliary cirrhosis. On the other hand, the SLP algorithm is superior in retrieving information in patients with primary biliary cirrhosis with a recall value of 65%.