Implementation of K-Nearest Neigbors for Prediction of Motorcycle Service Waiting Times in Develop Tech

Abstract

This research examines the application of the K-Nearest Neighbors (KNN) algorithm to predict wait times at Develop Tech, a workshop that offers affordable services but has longer spare parts wait times compared to official workshops. The research method used is descriptive quantitative, involving data collection and processing. In this study, the KNN algorithm is utilized to predict wait times by measuring distances between data points using and performing voting based on the K value to determine the final prediction. Testing on 100 data points demonstrated that KNN could predict wait times very accurately, achieving 100% accuracy, precision, and recall at certain K values. The data was split into 80% for training and 20% for testing, a method commonly used in machine learning research to ensure a balance between training data and validation. The results indicate that the KNN algorithm is reliable for predicting wait times with optimal performance at K values between 3 and 10. These findings support the conclusion that the KNN algorithm functions effectively in predicting wait times at the workshop.