Structured Airport Typology: A Layered Approach to Classifying Airport Types

Анотація

A two-stage multi-class classification method for recognizing airport types is developed. The method involves sequential binary clustering that optimizes the order in which classes are compared to improve the resulting classification metrics. Seven classifiers were used to achieve meaningful results. The method iteratively performs binary classification to determine the most effective sequence for distinguishing classes. Initially, a binary classification was performed to distinguish international airports from non-international airports. Subsequently, another binary classification was performed to distinguish regional airports from airfields. Numerical experiments were conducted and the F1 score was calculated to evaluate the performance. The minimum representative volume for the training dataset was determined and feature extraction was completed. The developed optimization method was verified and all metrics were cross-validated. The method was applied to recognize and classify international airports, regional airports, and airfields.

Класифікація

Ідентифікатори

Схожі роботи