Rule-Based Inference Model for Automated Sakip Performance Report Validation at RSUD XYZ

Abstract

The Government Agency Performance Accountability System (SAKIP) is an important instrument for ensuring that government programs can be accounted for based on measurable performance achievements. However, performance report validation at RSUD XYZ is still performed manually, requiring considerable validator time and attention and potentially producing inconsistent judgments. This study aims to develop a rule-based inference model for automated SAKIP performance report validation. The main contribution is a two-stage inference model that integrates the transformation of target and realization data into performance facts with Rule-Based Reasoning using Forward Chaining. The first stage calculates performance achievement according to positive and negative indicator characteristics, while the second stage uses the resulting facts as working memory and matches them against IF–THEN rules. The study employed a Research and Development approach with the Waterfall development model. Data were collected through observation, interviews, document analysis, and 20 performance indicator datasets validated by three validators. The results showed 100% conformity between system decisions and validator assessments with a 0% gap. The average validation time decreased from 6.78 seconds manually to 1.00 second automatically, representing an 85.25% improvement in time efficiency. Software quality evaluation produced Functional Suitability of 98.13%, Usability of 100%, Performance Efficiency of 84%, and Reliability of 96.53%. The proposed model demonstrates that separating performance-fact transformation from the inference process can support consistent and traceable validation.