Penerapan Hybrid Naïve Bayes dan Decision Tree dalam Sistem Pakar Diagnosis Penyakit Mulut Berbasis Android

Abstract

Oral diseases such as stomatitis, gingivitis, and candidiasis are often undetected at an early stage due to limited public knowledge and restricted access to healthcare services, leading to delays in treatment. This study aims to develop an Android-based expert system for diagnosing oral diseases using a hybrid approach that combines the Naïve Bayes and Decision Tree algorithms. The Naïve Bayes method is applied to calculate the probability of symptoms associated with potential diseases, while the Decision Tree generates diagnostic rules that are more transparent and interpretable. The research stages include a literature review, data collection and validation of disease symptoms from medical experts, development of the hybrid model, implementation into an Android application, and system testing using cross-validation and a confusion matrix. The expected outcomes include a prototype Android application for oral disease diagnosis with a minimum accuracy of 85%, a scientific article published in a nationally accredited journal indexed in Sinta, and additional outputs such as copyright registration of the application and publication of a book. This study is expected to improve public access to early diagnosis of oral diseases, support early detection, and contribute to the advancement of digital health systems based on artificial intelligence in Indonesia.