Performance Prediction Model for Predictive Maintenance Based on K-Nearest Neighbor Method for Air Navigation Equipment Facilities

The aging of air navigation facilities leads to a decline in equipment performance and reliability, posing significant challenges for aviation organizations in the large-scale replacement of outdated systems, as well as in managing budgets and resources. To address these issues, a performance prediction model for equipment was developed using the K-Nearest Neighbor (KNN) method, aimed at enhancing maintenance planning and budget management. The development process includes collecting damage report data from facilities, preprocessing the data to ensure quality and consistency, and applying the KNN algorithm to generate accurate predictions. The KNN model, with the parameter n_neighbors = 2, achieved a high accuracy of 89.13% on the test data, with the best performance in class 1 classification. These results demonstrate the superiority of KNN over other models, such as Random Forest, which achieved an accuracy of 77%, and Logistic Regression, which only reached 41%. This research not only validates the effectiveness of the KNN model in predicting the performance of air navigation equipment facilities but also contributes significantly to maintenance efficiency. By using the KNN method, aviation organizations can plan maintenance more proactively and efficiently, minimizing the risk of unexpected failures. Moreover, the model aids in more effective budget preparation by adjusting maintenance priorities according to the specific needs and conditions of the facilities. This research focuses on providing a practical and reliable solution for maintenance planning and improved budget management, ultimately enhancing the performance and safety of flight services.