Personalized Hotel Recommender System Based on Graded Logic with Asymmetric Criteria

Abstract

Personalizing hotel recommendations for tourists is challenging because it needs to model a logical aggregation of multiple criteria in terms of the user’s decision process and preferences. In this paper, we propose a hierarchical multi-criteria hotel evaluation model that includes important hotel properties as well as hotel ratings by previous customers. The recommender system proposed, called HR.LSP, applies different graded logic aggregation operators following the method of Logic Scoring of Preference (LSP). Asymmetric aggregation is used for some of the criteria, where the user can not only define the importance of inputs, but also indicate if they are mandatory or optional. This is the first recommender system based on graded logic and the first non-professional use of LSP. The application of graded logic operators significantly improves the decision-making process of hotel selection for each user. The case of Riohacha (an emerging tourist destination in Colombia) is studied. In addition to quality and location criteria, some security information is included in the model as it is of utmost importance to provide to travelers a secure experience at the destination.