Sistem Pakar Diagnosa Awal Cacar Monyet Menggunakan Logika Fuzzy Metode Tsukamoto dan Mesin Inferensi Forward Chaining Berbasis Android

Abstract

Rising monkeypox cases in Indonesia require accessible early detection systems given conventional diagnosis limitations and lack of public understanding about disease symptoms. This research develops an early diagnosis expert system for monkeypox by integrating Fuzzy Tsukamoto and Forward Chaining methods based on responsive web. Iterative and Incremental Development approach was applied through three development iterations. Forward Chaining traces 7 discrete symptoms (G01-G07) producing qualitative diagnosis which is then reinforced by Fuzzy Tsukamoto through fuzzification of 3 continuous variables (body temperature, rash count, lymph node swelling) with 27 inference rules and weighted average defuzzification to quantify risk level. Integration mechanism works in layers where Forward Chaining verifies binary symptom completeness as initial diagnosis, subsequently the qualitative output is reinforced with quantitative certainty value (Z) from Tsukamoto as risk stratification, producing comprehensive diagnosis that overcomes single-method system limitations. Testing on 20 cases showed clear separation between low-risk (Z: 0.331-0.463) and high-risk (Z: 0.814-0.990) categories. Validation using medical expert diagnosis as gold standard yielded 95% accuracy, 90% sensitivity, 100% specificity, 100% precision, and 94.7% F1-score, proving system capability in accurate diagnosis. Black Box testing validated all system functionalities running error-free. Layered integration of both methods proved effective in producing objective diagnosis with high accuracy, although further research is needed for multi-disease base expansion and validation using actual clinical data from healthcare facilities.