Perceptual maps of Turkish airline services for different periods using supervised machine learning approach and multidimensional scaling

Abstract

In the airline market, it is crucial for airline industry to determine the experiences, expectations and perceptions of passengers in order to apply positioning strategies on brands. In this study, we used 15,864 Turkish tweets sent to the official airline Twitter pages based in Turkey between 1st June and 1st September 2017. Then, we applied aspect-based sentiment analysis (ABSA) with supervised machine learning approach to classify tweets into airline service categories and sentiment polarity. Lastly, multidimensional scaling (MDS) employed to build perceptual maps of airline services for different periods. This study aims to explore how tweets reflect airline service quality attributes in perceptual maps for selected periods in Turkey. Our analysis shows that the perceptual positions of services change per period, which means that Twitter users perceived each service differently in each period. In terms of the importance of airline service quality attributes website services, convenience of flight, and in-flight entertainment were the most critical disparities perceived by users compared to other attributes considering in the periods being examined.