Perancangan Sistem Monitoring Kualitas Madu Ber- Basis IoT dengan Metode K-Nearest Neighbors (K-NN) Berdasarkan Sumber Makanan dan Lingkungan Peternakan Lebah

Abstract

Honey quality is a key factor in the beekeeping industry that requires continuous Monitoring to ensure optimal product standards. This research aims to develop an Internet of Things (IoT) based honey quality Monitoring system using the K-Nearest Neighbors (K-NN) algorithm to classify Trigona honey quality based on apiary environmental parameters and bee food sources. The system utilizes ESP32 microcontroller integrated with six sensors including DHT22 for temperature and humidity, pH sensor SKU:SEN0161, MQ-135 gas sensor, TDS sensor, and LDR sensor for real-time Monitoring. Sensor data is transmitted to Firebase cloud platform and processed using K-NN algorithm with K=3 value for honey quality classification. Testing was conducted on 315 real data from three types of Trigona honey: Biroi Honey, Itama Honey, and Laeviceps Honey. Evaluation results using confusion matrix show the system achieves 88.25% accuracy with highest precision of 92.70% for "Good Quality" category and 89.60% recall for "Excellent Quality" category. TDS and pH parameters proved to be the most significant factors in classification with 85-92% confidence level. The developed web dashboard is capable of presenting real-time data visualization and quality analysis with automatic alert system. This system proves effective for practical implementation in honey quality Monitoring at Trigona bee farms and can assist beekeepers in maintaining consistency of honey product quality.