Infection Disease Classification in Indonesian Language using Machine Learning

Infectious diseases are still the main cause of death worldwide due to their dangerous and fast-spreading natures. One way to diagnose infectious diseases quickly and accurately is to create a machine learning model that aims to assist young doctors or health practitioners. In this study, we used the support vector machine, k-nearest neighbors, and the decision tree algorithm to classify types of infectious diseases based on the similarity of signs and symptoms. These methods were trained on documents in the Indonesian Language, which has its weight extracted using the Term Frequency-Inverse Document Frequency (TF-IDF). Model training and testing were carried out using 464 data with 20 classes of infectious diseases. Test results in this study indicate that support vector machine using stopword NLTK with unigram, bigram, and trigram get anaccuracy of 87%, 81%, and 76%, followed by the K-nearest neighbors using stopword NLTK get an accuracy of 83%, 81%, and 77%. The decision tree using stopword NLTK gets an accuracy of 66%, 60%, and 53%, scoring the lowest compared to the other two.