Tez Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: Eskişehir Teknik Üniversitesi, Lisansüstü Eğitim Enstitüsü, Havacılık Yönetimi, Türkiye
Tez Danışmanı: Dr. Öğr. Üyesi Emircan Özdemir
Tezin Onay Tarihi: 2026
Tezin Dili: Türkçe
Özet:
Intense competition in the aviation industry makes it essential for airlines to measure the criteria that determine passenger satisfaction in an objective and reliable manner. The aim of this study is to analyze airline passenger satisfaction using a hybrid approach that integrates machine learning and multi criteria decision making (MCDM) methods, and to comparatively evaluate the service quality performance of 18 leading European airlines. In this study, which utilized 1,927 passenger reviews obtained from the Skytrax platform, the final importance weights of six service quality criteria were determined using the arithmetic mean of weights derived from a multi layer perceptron (MLP) and SHAP based explainable artificial intelligence integration, whose stability was verified using the entropy method with 1,000 independent repetitions. The resulting hybrid weights were fed into the RAWEC method to rank the alternatives; the methodological reliability of the results was demonstrated through comparative sensitivity and correlation analyses conducted using the VIKOR, TOPSIS, WASPAS, CoCoSo, MABAC, MARCOS, and COPRAS methods. The findings reveal that ground services are the most influential factor in determining passenger satisfaction and that Aegean Airlines demonstrated the highest service quality performance, thereby offering a methodological and practical original contribution to the literature on machine learning and MCDM. By integrating entropy and SHAP based weighting approaches, this study offers both a methodological and practical contribution to the literature on machine learning and MCDM.