Article: Analysis of tourist emotions and behaviour patterns using deep learning Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.8 pp.53 - 73 Abstract: This study explores the emotions and behavioural patterns of travellers using advanced deep learning techniques. A hybrid model combining CNN and LSTM networks was developed to analyse Twitter data related to travel in Thailand, enabling the identification of key emotional states such as joy, surprise, fear and melancholy. The proposed approach provides valuable insights for recommendation systems and tourism management, as tourists' emotions significantly influence travel decisions and satisfaction levels. By analysing emotional tendencies, tourism services can enhance the overall visitor experience. While previous research has largely relied on machine learning and lexicon-based methods for textual emotion detection, recent advancements in deep learning have demonstrated superior predictive accuracy for sentiment and behavioural analysis. In this study, CNN-LSTM models, complemented by feature extraction techniques using DenseNet and AlexNet, were employed. The hybrid model achieved 91% accuracy, surpassing conventional methods, with joy and surprise being the most accurately classified positive emotions. Inderscience Publishers - linking academia, business and industry through research

Title: Analysis of tourist emotions and behaviour patterns using deep learning

Authors: Limin Wang

Addresses: School of Management, Zhengzhou University of Industrial Technology, Zhengzhou, Henan, 450000, China

Abstract: This study explores the emotions and behavioural patterns of travellers using advanced deep learning techniques. A hybrid model combining CNN and LSTM networks was developed to analyse Twitter data related to travel in Thailand, enabling the identification of key emotional states such as joy, surprise, fear and melancholy. The proposed approach provides valuable insights for recommendation systems and tourism management, as tourists' emotions significantly influence travel decisions and satisfaction levels. By analysing emotional tendencies, tourism services can enhance the overall visitor experience. While previous research has largely relied on machine learning and lexicon-based methods for textual emotion detection, recent advancements in deep learning have demonstrated superior predictive accuracy for sentiment and behavioural analysis. In this study, CNN-LSTM models, complemented by feature extraction techniques using DenseNet and AlexNet, were employed. The hybrid model achieved 91% accuracy, surpassing conventional methods, with joy and surprise being the most accurately classified positive emotions.

Keywords: tourist emotions; behaviour patterns; emotion recognition; tourist experience analysis; deep learning; behavioural analytics; artificial intelligence in tourism.

DOI: 10.1504/IJICT.2026.151598

International Journal of Information and Communication Technology, 2026 Vol.27 No.8, pp.53 - 73

Received: 26 Sep 2025
Accepted: 18 Oct 2025
Published online: 09 Feb 2026 *

Title: Analysis of tourist emotions and behaviour patterns using deep learning

Keep up-to-date