Expert System for Diagnosing Black Glutinous Rice Plant Diseases Using Forward Chaining and Backward Chaining Methods

Abstract

Black glutinous rice (Oryza sativa L. var. glutinosa) is a local Indonesian rice variety whose productivity is often reduced by pest and disease attacks, while accurate diagnosis remains difficult due to limited access to agricultural experts. This study develops a web-based expert system for diagnosing diseases in black glutinous rice plants by integrating Forward Chaining and Backward Chaining methods. The system was developed using the Expert System Development Life Cycle (ESDLC). Knowledge was acquired through interviews with the Puspamukti Farmers Group, experts from the Agrotechnology Study Program at Universitas Perjuangan Tasikmalaya, field observations, and literature review. The knowledge base consists of 30 symptoms, 6 pests and diseases, and IF–THEN production rules. In the proposed hybrid inference mechanism, Forward Chaining generates candidate diseases from user-selected symptoms, while Backward Chaining verifies the diagnosis before presenting treatment recommendations. System evaluation using Black Box Testing and expert validation on 20 test cases showed that all system functions operated correctly and achieved an accuracy of 95%, with 19 of 20 diagnoses matching expert assessments. The proposed hybrid approach provides a more consistent diagnostic process and can effectively support farmers in the early diagnosis and management of black glutinous rice diseases.