Detecting Novelty Seeking From Online Travel Reviews: A Deep Learning Approach
B. Venkata Sivaiah, N. Siva, N. Sunil Kumar, M. Sucharitha, Sachin Kumar, Y. Eswar · Advances in computer science research · 2024
An important source of experience-related data for comprehending novelty seeking (NS), a natural personality feature that affects travel motivation and location selection, is online travel reviews.Due to the large number and disorganization of reviews, manually categorizing them is difficult.Therefore, our aim is to develop a deep learning model and classification system to overcome these challenges.We propose a framework that combines four dimensions related to the NS personality trait and use a DL model called BERT-BiGRU to automatically identify NS in TripAdvisor reviews using a dataset of 30,000 reviews.The classifier using the NS multi-dimensional scale and the BERT-BiGRU multi-dimensional scale accurately identified the NS personality trait.It achieved high accuracy and F1 scores.The BERT-BiGRU model outperformed other DL models in terms of accuracy.This study shows how computer methods can be used to automatically determine personality traits from travel reviews.It provides a comprehensive framework for categorizing personality traits in order to benefit marketing and recommendation systems in the travel industry.