An Expert System Using Ontology as Knowledge Base for Personalized Rice Cultivation Suggestion

This work presents a framework to assist Thai rice farmers in cultivation planning. Knowledge of rice production is extracted from best practice documents and kept in ontology. Concepts relevant to rice production such as soil type of every location in Thailand and rice seeds with their strong and weak properties are semantically drawn in the ontology. The ontology associated with rules to make decision is exploited as a knowledge base for the system to infer a reasonable suggestion based on farmers' specific information, context and need. According to farmers input, different suggestions can be given including seed selection, land preparation, concern in cultivation based on preferred seed, etc. From testing, the generated recommendations were satisfied by rice production experts for precise criteria analysis and proper reason selection in inference process. An experiment result of seed selection showed that the system generated suggestions with 0.56 precision score and 0.97 recall score comparing to human expert suggestions.