Sistem Prediksi Efisiensi Konsumsi Listrik Konsumsi Listrik Rumah Tangga dengan Fuzzy Logic

Wicaksono, Arief Tri
Syahrorini, Syamsudduha
Anshory, Izza
Jamaaluddin, Jamaaluddin

TIN: TERAPAN INFORMATIKA NUSANTARA
Vol 6 No 6 (2025): November 2025

The increase in household electricity consumption in Indonesia poses a major challenge to efficient and sustainable energy management. This research aims to develop a household electricity consumption status prediction system based on Mamdani fuzzy logic using the Python programming language. The system was designed to classify electricity usage into “efficient” or “inefficient” categories based on three main input variables: electrical power, usage duration, and number of electronic devices. The design process included fuzzification, rule base construction, Mamdani inference, and defuzzification, with the final results integrated into Google Sheets. System validation was performed by comparing 55 predicted results to actual labeled data. Performance evaluation using confusion matrix analysis showed an overall accuracy of 80% (44 out of 55 data) and a Mean Squared Error (MSE) value of 0.20. The main advantage of this model is its perfect sensitivity (recall) (100%) in detecting inefficient classes, as evidenced by a False Negative (FN) value of 0. The system successfully identified all 39 actual inefficient cases, although there were 11 efficiency cases that were classified as inefficient (False Positive). This research contributes to providing a precise digital tool that can increase public awareness of energy efficiency and support the control of household electricity consumption.