K-Nearest Neighbor (K-NN) Method for Disturbance Classification of Customer Wifi Networks at PT. Global Karya Wanda

Abstrak

This study aims to develop and implement a Wi-Fi network disturbance classification system using the K-Nearest Neighbor (K-NN) algorithm at PT. Global Karya Wanda. The purpose of this research is to identify and classify Wi-Fi network conditions based on standard categories such as interference, troubleshooting, disconnection, and signal loss, thereby improving the efficiency and accuracy of network monitoring. The system was designed and developed using PHP and MySQL, with datasets obtained from PT. Global Karya Wanda’s operational network records. The classification process employed the K-NN algorithm to distinguish between Disturbance and Not Disturbance network states. The experimental results demonstrate that the K-NN method provides fast, automatic, and accurate classification performance, supporting the company in optimizing its troubleshooting workflow and enhancing customer service reliability. From a practical standpoint, the model enables more systematic network performance monitoring and proactive disturbance management. Scientifically, this research contributes to the application of machine learning algorithms in network performance analysis and telecommunications service optimization. Future studies are recommended to integrate hybrid approaches such as KNN–SVM or machine learning API integration to improve classification accuracy, scalability, and real-time responsiveness in larger and more dynamic network environments.