Mobile Social Travel Recommender System
Abstract
Travel Recommender Systems (TRSs) help tourists discovering and selecting the Points of Interest (POIs) that best fit their preferences. Recommendations rely on the data available about the POIs of a destination, the knowledge about tourists and their preferences about categories, and recommendation algorithms. This paper presents a Mobile Social TRS. The recommendation process is divided in two independent processes: the generation of user models and the calculation of the recommended POIs. The recommender generates user models taking into account their explicit preferences about categories, demographic information, and the tags they have created. Then, similarities between users are based on the POIs they have rated. Finally, a hybrid filtering algorithm combines these models with a content-based and a collaborative filtering algorithm to calculate a list of recommended POIs. The recommender has been integrated in a mobile prototype of the CRUMBS social network and preliminary results of its partial validation are presented.
BIB_text
author = {Ander Garcia, Isabel Torre, María Teresa Linaza},
title = {Mobile Social Travel Recommender System},
pages = {3-16},
keywds = {
Travel recommender system, User model, Hybrid recommendation algorithm, Mobile
}
abstract = {
Travel Recommender Systems (TRSs) help tourists discovering and selecting the Points of Interest (POIs) that best fit their preferences. Recommendations rely on the data available about the POIs of a destination, the knowledge about tourists and their preferences about categories, and recommendation algorithms. This paper presents a Mobile Social TRS. The recommendation process is divided in two independent processes: the generation of user models and the calculation of the recommended POIs. The recommender generates user models taking into account their explicit preferences about categories, demographic information, and the tags they have created. Then, similarities between users are based on the POIs they have rated. Finally, a hybrid filtering algorithm combines these models with a content-based and a collaborative filtering algorithm to calculate a list of recommended POIs. The recommender has been integrated in a mobile prototype of the CRUMBS social network and preliminary results of its partial validation are presented.
}
date = {2014-01-21},
year = {2014},
}
