Feeling a destination through the “right” photos: A machine learning model for DMOs’ photo selection

• Published in: Tourism management (1982) Vol. 65; pp. 267 - 278
• Main Authors: Deng, Ning, Li, Xiang (Robert)
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.04.2018
• Subjects: Data mining; Destination image; Flickr; Photo recommendation; UGC photos; UGC photos; Flickr; Data mining; Destination image; Photo recommendation
• ISSN: 0261-5177; 1879-3193
• DOI: 10.1016/j.tourman.2017.09.010

Photos are important carriers in destination image communication. Currently, efficiently selecting appropriate photos for destination promotion remains a major challenge for DMOs, a problem closely related to the discrepancy between projected and received destination images. During the photo selection process, contents that can best evoke viewers' potential motives should be considered favorably. This project proposes and implements a machine learning-based model to assist DMOs with photo content selection. The proposed protocol ranks candidate photos describing a specific theme from viewers’ perspective. In the present empirical study, over 20,000 Flickr photos of New York City taken by foreign tourists were analyzed to demonstrate the effectiveness of this approach. The results indicate that the proposed method can facilitate the selection of destination photos and address the pronounced gap between projected and received images. •DMO should take UGC into consideration when projecting destination image in online era.•A machine learning model to recommend suitable advertising photos for DMO, based on UGC photos.•The naive Bayesian classifier is adopted to rank the photos according to their relevancies.•Flickr's metadata and comments are applied to build a mapping relation between cognitive and affective image.•The advantage of the approach is to shrink the gap between projected and received image.