KLASIFIKASI PENYAKIT KULIT KUCING BERBASIS GEJALA KLINIS MENGGUNAKAN SVM DAN CHI-SQUARE

Abstract

Scabies and dermatophytosis in cats often present overlapping clinical symptoms, making initial screening difficult. This study aimed to develop a web-based application for classifying feline skin diseases using the Support Vector Machine (SVM) algorithm. Medical record data from an Animal Health Center were preprocessed, transformed into binary features, and selected using the Chi-Square test. Two classification scenarios were developed: Mode 1 distinguished scabies from non-scabies using 6,864 records, while Mode 2 differentiated scabies from dermatophytosis using 752 records. The results showed that Mode 1 achieved 86.1% accuracy, 91% recall, and a ROC AUC of 0.903, whereas Mode 2 achieved 70.1% accuracy, 83% scabies recall, an F1-score of 0.69, and a ROC AUC of 0.715. The Chi-Square analysis identified crusts or scabs as one of the most influential features for scabies classification. The novelty of this study lies in combining tabular clinical symptoms, Chi-Square feature selection, two SVM classification scenarios, and web-based implementation. Black Box Testing confirmed that all primary functions operated as designed. The system can support screening and veterinary decision-making but should not replace definitive clinical diagnosis.