Ontology-Based Natural Language Processing for Thai Herbs and Thai Traditional Medicine Recommendation System Supporting Health Care and Treatments (THMRS)

Herbs become increasingly popular in health care and treatments these days, because they are safe under careful usages. Although Thai herbs information can be retrieved frommanywebsites, it takes long searching times for discovering the information that matches with users' requirements. Therefore, this research aimed to develop a Thai herbs and Thai traditional medicines recommendation system for health care and treatments using ontology-based natural language processing (THMRS). In addition, users can use the proposed recommendation system via web or Windows application. The proposed THMRS is composed of ontology-based databases called Thaiherb, searching process using natural language combined with word tokenization and spell checking by ISG (Index of Similarity Group) Algorithm, translation from natural language to SPARQL commands is designed for semantic search, and the decision making whether information found will be delivered to users is done by co-occurrence density analysis. The search result was displayed in a single website or a window which is fast and user-friendly. The experiments show that the average F-Measure is 95.07%, the average precision is 93.76%, and the average recall is 96.42% which indicate that THMRS is very high efficiency.