<![CDATA[This study introduces a novel integration of the Mamdani Fuzzy Inference System (FIS) with a web-based recruitment platform for Special Job Fair (BKK) institutions at vocational schools a context not previously addressed in the fuzzy decision-support literature. The current manual process at BKK Jaya Abadi SMKN 1 Bukateja relies on paper-based workflows, causing data loss risks, prolonged processing times, and subjective multi-criteria evaluation. The proposed system uses three input variables psychotest, interview, and health scores processed through fuzzification, an 18-rule inference base, MAX aggregation, and centroid defuzzification to generate objective recruitment recommendations. Testing on 35 applicant records yielded MAPE = 4.29% and accuracy = 95.71%, outperforming both the manual assessment baseline (MAPE = 12.86%, accuracy = 87.14%) and a simple weighted sum alternative (MAPE = 8.42%, accuracy = 91.58%). A paired t-test (α = 0.05) confirmed that accuracy differences are statistically significant. According to Lewis (1982), the system falls in the "Very Good" accuracy category. The system significantly reduces subjective inconsistencies in candidate evaluation while maintaining transparency and alignment with expert judgment, offering a practical and interpretable approach to automating multi-criteria recruitment decisions in BKK environments. ]]>
Copyrights © 2026
