Transforming Customer Experience in the Airline Industry: A Comprehensive Analysis of Twitter Sentiments Using Machine Learning and Association Rule Mining
Maliha Tayaba, Eftekhar Hossain Ayon, Md Tuhin Mia, Malay Sarkar, Rejon Kumar Ray, Md Salim Chowdhury, Md. Ibrahim Khalil Al-Imran, Nur Nobe, Bishnu Padh Ghosh, MD Tanvir Islam, Aisharyja Roy Puja · Journal of Computer Science and Technology Studies · 2023
The airline industry places significant emphasis on improving customer experience, and Twitter has emerged as a key platform for passengers to share their opinions. This research introduces a machine learning approach to analyze tweets and enhance customer experience. Features are extracted from tweets using both the Glove dictionary and n-gram methods for word embedding. The study explores various artificial neural network (ANN) architectures and Support Vector Machines (SVM) to create a classification model for categorizing tweets into positive and negative sentiments. Additionally, a Convolutional Neural Network (CNN) is developed for tweet classification, and its performance is compared with the most accurate model identified among SVM and multiple ANN architectures. The results indicate that the CNN model surpasses the SVM and ANN models. To provide further insights, association rule mining is applied to different tweet categories, revealing connections with sentiment categories. These findings offer valuable information to help airline industries refine and enhance their customer experience strategies.