Penerapan Algoritma K-Nearest Neighbor (K-NN) untuk Klasifikasi Status Monitoring Automatic Pump Water Machine Studi Kasus: Industri Manufaktur

Indra, Indra Mora
Adlian, Adlian Jefiza

Technology Sciences Insights Journal
Vol. 2 No. 2 (2025): Technology Sciences Insights Journal

In the modern industrial world, real-time monitoring of system conditions is  crucial to maintain efficiency and prevent equipment damage. This research aims to classify industrial system conditions based on sensor data using the K Nearest Neighbors (KNN) algorithm. The data used consists of four main  parameters namely pressure, flow rate, voltage, and engine speed (RPM),  which are then classified into three conditions: Alert, Critical, and Normal.  Preprocessing is done with Min-Max normalization and division of data into  training and test data. The evaluation results show that the KNN method is  able to achieve an accuracy of 58% with a mean squared error (MSE) value of 1.06 and an average cross-validation accuracy of 64%. These results show that  KNN is effective enough to be used as an initial method for industrial system  condition detection, although the classification performance for the Critical  category can still be improved.