Analysis of Airline Passenger Satisfaction Using Decision Tree and Naïve Bayes Algorithms

Abstract

In the dynamic landscape of airline services, comprehending the intricacies that mold customer satisfaction is paramount to elevating overall service quality. This study aspires to dissect these pivotal elements, contributing nuanced insights that can propel the enhancement of customer satisfaction within the industry. A multifaceted investigation encompasses analyzing demographic data, exploring underlying factors significantly shaping passenger satisfaction, and identifying the most adept model for forecasting imminent passenger satisfaction outcomes. A model was meticulously crafted by leveraging a decision tree algorithm to discern the substantial variables influencing passenger satisfaction. Simultaneously, the Naïve Bayes algorithm was harnessed to prognosticate forthcoming passenger satisfaction. The findings underscore the diverse facets of the flying experience impacting satisfaction, with both ctree and rpart decision tree algorithms spotlighting critical factors, such as online boarding, inflight entertainment, WiFi service, class, and travel type. The Naïve Bayes algorithm demonstrates around 87% accuracy in predicting passenger satisfaction, underscoring its efficacy in discerning patterns within this complex realm.