Decoding customer experiences in rail transport service: application of hybrid sentiment analysis

Document Type

Journal Article

Publication Date


Subject Area

mode - rail, place - asia, planning - methods, planning - service improvement, technology - passenger information


Sentiment analysis, Machine learning, Naïve Bayes, Twitter, Customer experience, Transportation, Rail


The paper aims to enhance customers’ satisfaction levels by identifying improvements in the service quality of the rail transport industry in developing countries such as India. A multi-algorithmic combination of a LEXICON analysis and a Naïve Bayes machine learning hybrid approach to sentiment analysis is performed for identifying passengers’ opinions on the services provided by Indian Railways. Inputs were gathered from the Twitter microblogging platform. Data analysis reveals that the ticket reservation and refund process, delay in operational activities, and abhorrent behavior of staff were crucial areas in which Indian Railway service needs improvement. The study imparts a conceptual methodology/process for implementing a hybrid multi-algorithmic LEXICON and machine learning techniques in sentiment analysis. The model proves to take less time to process, train, and test data than stand-alone LEXICON or machine learning-based approaches. Managers, practitioners, and researchers may use this approach to understand customer experience especially in rail transportation but also across hospitality sectors such as hotels, restaurants, education, and hospitals.


Permission to publish the abstract has been given by SpringerLink, copyright remains with them.