Abstract:
Airlines must understand customer feelings and preferences to stay competitive and
successful in today's dynamic and fiercely competitive airline market. Customer reviews
provide extensive information about passengers' experiences, attitudes, and expectations.
Analysing these evaluations using modern machine learning (ML) techniques allows
airlines to extract actionable insight from massive amounts of unstructured data to make
informed decisions and better serve customers. This study analyses Bangladeshi airline
customer reviews using boosting and voting ML. This study uses 3000 reviews from
official airline websites like Novo Air, Bangladesh Biman, and US Bangla and social media
traveller groups to understand passenger opinions of Bangladeshi airlines' services. This
research uses advanced ML algorithms like CatBoosting, AdaBoosting, XGBoosting, and
Gradient Boosting and ensemble techniques like Voting RM, SVM, and Naive Bayes to
find hidden patterns, trends, and insights in customer feedback. highest accuracy 87%
achieved by SVM and random-forest voting techniques.
