Let pictures speak: hotel selection-recommendation method with cognitive image attribute-enhanced knowledge graphs

Purpose: The purpose of this paper is to study the importance of image data in hotel selection-recommendation using different types of cognitive features and to explore whether there are reinforcing effects among these cognitive features. Design/methodology/approach: This study represents user-generated images “cognitive” in a knowledge graph through multidimensional (shallow, middle and deep) analysis. This approach highlights the clustering of hotel destination imagery. Findings: This study develops a novel hotel selection-recommendation model based on image sentiment and attribute representation within the construction of a knowledge graph. Furthermore, the experimental results show an enhanced effect between different types of cognitive features and hotel selection-recommendation. Practical implications: This study enhances hotel recommendation accuracy and user satisfaction by incorporating cognitive and emotional image attributes into knowledge graphs using advanced machine learning and computer vision techniques. Social implications: This study advances the understanding of user-generated images’ impact on hotel selection, helping users make better decisions and enabling marketers to understand users’ preferences and trends. Originality/value: This research is one of the first to propose a new method for exploring the cognitive dimensions of hotel image data. Furthermore, multi-dimensional cognitive features can effectively enhance the selection-recommendation process, and the authors have proposed a novel hotel selection-recommendation model.