A Comparison of Semantic and Ontology for Searching Disease-Related QA Answers from Online Health Consultation Texts
Safitri Juanita, Arya Nur Razzaq, Muhammad Raihan, Diana Purwitasari, I Ketut Eddy Purnama, Mauridhi Hery Purnomo · 2024
The inability of the search feature to find relevant information based on keywords in the set of questions (Q) and answers (A) in online health consultation (OHC) may lead to inaccurate search results. Therefore, it is necessary to investigate the best method to help search engines find relevant information according to keywords, mainly Q&A containing medical text. This study aims to find a method that is significantly accurate in finding relevant information related to diseases based on the history Q&A of OHC. This study compares two methods, namely semantics and ontology. Both methods were tested, and their performance results were compared using a prototype website for searching disease-related Q.A. text that we developed. The research stages were divided into four parts, namely data selection, back-end scenarios, front-end and testing, with the research dataset in the form of a collection of doctor’s answer texts from one of the OHCs with the topic of high-risk diseases according to WHO, namely Tuberculosis (T.B.), Diarrhea, HIV/AIDS, and Kidney Stones. The results show that the ontology method is more accurate in finding disease information related to questions and answers on OHC, with a precision of 75%, recall of 100%, and F1-score of 82% compared to the semantic method.