Diagnosis of Feline Skin Disease Using C4.5 Algorithm

Abstract

Cats are one type of animal that is very popular with many people, and there is even a community of cat fans known as cat lovers. The health indicator in cats lies in the condition of their skin, so it needs special care to maintain their skin condition. Many cat owners are not aware of the skin diseases suffered by their cats. This is due to the owner's limited knowledge of the diseases experienced by cats and the difficulty in identifying the similar symptoms experienced by cats. To overcome this problem, we need a method to diagnose skin diseases that occur in cats. Diagnosis of symptoms of cat skin disease can be done by a classification method in data mining. In this study, the classification method used to diagnose skin diseases in cats is the C4.5 algorithm. The dataset used was obtained from the animal clinic “Purple Shop” in Malang. The algorithm testing process is carried out using k-fold cross-validation. Algorithm performance evaluation is measured by using a confusion matrix, namely by measuring the value of accuracy, precision, and recall. The results of this study indicate that the resulting accuracy value is 95.42%, the average precision is 96.93%, and the average recall is 97.19%. These results indicate that the C4.5 algorithm shows a very high level of performance and can be applied to diagnose symptoms of skin disease in cats.