Ranking hotel products by integrating online ratings from multiple platforms

Abstract

The proliferation of third-party platforms has led to the same product or service appearing across multiple platforms. To facilitate consumers' purchase decisions, it is essential to rank products based on online ratings from various platforms. However, ranking such products poses challenges due to discrepancies across platforms. In this paper, we propose a model for ranking products based on the evidential reasoning approach. The proposed model aims to overcome these challenges by determining a finite set of possible hypotheses, with the power set containing all possible subsets and a basic probability assignment (BPA) based on the distribution of ratings on a given platform. The model then calculates the weight of each platform and adjusts the BPA using the importance discounting method. It combines discounted BPAs using the proportional conflict redistribution rule number 5. The belief structure is then transferred into a score to rank alternatives. Finally, we validate our model by ranking hotels in Hong Kong, China, collected from popular platforms such as TripAdvisor, Agoda, Booking.com, Expedia, and Trip.com. Our case study demonstrates that our model leverages evidence combination to neutralize inconsistent information across platforms and maintain consistent opinions.