Exploring cross-cultural disparities in tourists' perceived images: a text mining and sentiment analysis study using LDA and BERT-BILSTM models

Qiuying Chen,Ronghui Liu,Qingquan Jiang,Shangyue Xu
DOI: https://doi.org/10.1108/dta-10-2023-0645
IF: 1.713
2024-03-22
Data Technologies and Applications
Abstract:Purpose Tourists with different cultural backgrounds think and behave differently. Accurately capturing and correctly understanding cultural differences will help tourist destinations in product/service planning, marketing communication and attracting and retaining tourists. This research employs Hofstede's cultural dimensions theory to analyse the variations in destination image perceptions of Chinese-speaking and English-speaking tourists to Xiamen, a prominent tourist attraction in China. Design/methodology/approach The evaluation utilizes a two-stage approach, incorporating LDA and BERT-BILSTM models. By leveraging text mining, sentiment analysis and t-tests, this research investigates the variations in tourists' perceptions of Xiamen across different cultures. Findings The results reveal that cultural disparities significantly impact tourists' perceived image of Xiamen, particularly regarding their preferences for renowned tourist destinations and the factors influencing their travel experience. Originality/value This research pioneers applying natural language processing methods and machine learning techniques to affirm the substantial differences in the perceptions of tourist destinations among Chinese-speaking and English-speaking tourists based on Hofstede's cultural theory. The findings furnish theoretical insights for destination marketing organizations to target diverse cultural tourists through precise marketing strategies and illuminate the practical application of Hofstede's cultural theory in tourism and hospitality.
computer science, information systems,information science & library science
What problem does this paper attempt to address?