Shared Accommodation Recommendation Based On Bidirectional Selection And Review Mechanisms

Recommendation system is widely applied in accommodation to improve the booking efficiency of users.Recently,shared accommodation that attracts local residents to rent out houses has shown rapidly popular.Research indicates that guests value the host attribute when booking a room and hosts have the right to and practically decline to reservation request.Current accommodation recommendation methods would generate invalid suggestions for neglecting such factors,we thus propose a two-stage recommendation model,SARBSR(Shared Accommodation Recommendation based on Bidirectional Selection and Review mechanisms),and intend to recommend guests rooms that not only meet guests demand but also ensure guests will probabaly receive the hosts' approving.In first stage,the host attribute is quantified by doing sentiment analysis(SA)of reviews about the host and integrated into rooms' multiattribute rating.Based on the rooms' multi-attribute rating containg hosts' credibility,Multiattribute rating recommendation(HMAR)and content-based collaborative filtering(HCBCF)methods are implemented to generate a Candidate list.Noted that LDA is used to analyze users' multi-preference from reviews.Based on recommendation candidate list,we assess hosts' acceptance degree of guests via trust evaluation.SVR is adopted to model hosts' trust evaluation with the characteristic of potential guests.Based on the real-world data from Airbnb,we demonstrate that considere the host attribute affecting guests' decision making and hosts' selection of guests improve recommendation efficiency,respectively.Experiment results also indicate that SARBSR outperforms User-based Collaborative filtering(CF),Matrix Factorizing(MF),multi-attribute rating recommendation(MAR)and content based collaborative foltering(CBCF)in terms of top-k recommendation.Furthermore,SARBSR expresses application significance in increasing the transaction efficiency of guests and the revenue of operators that charge service fee of transactions.