Accommodation Recommendation on Shared Platforms Considering Bidirectional Selection and Review Mechanisms

Abstract

In recent years, shared accommodation platforms have developed rapidly and become increasingly popular. They commonly adopt unique bidirectional selection and review mechanisms. However, most existing accommodation recommendation strategies overlook the impact of these mechanisms on the recommendation performance. To address this gap, we propose a two-stage recommendation method as Shared Accommodation Recommendation based on Bidirectional Selection and Review mechanisms (SARBSR) to improve the recommendation performance of shared accommodation. The first stage combines a multi-attribute recommendation method (MAR) with a hybrid recommendation method based on similarity and ratings (HRSR) to predict guests’ preferences and generate candidate lists. In the second stage, considering bidirectional selection and reviews mechanisms, we assess hosts’ willingness to accept guests’ requests via trust evaluation and select rooms from the candidate lists as the final recommendation strategy. To evaluate the performance of SARBSR, an empirical study on real-world data from Airbnb is conducted. The results demonstrate the validity of SARBSR and indicate the necessity of considering bidirectional selection and review mechanisms.