Recommending POIs Based on the User’s Context and Intentions

Abstract

This paper describes a Recommender System that implements a Multiagent System for making personalised context and intention-aware recommendations of Points of Interest (POIs). A twoparted agent architecture was used, with an agent responsible for gathering POIs from a location-based service, and a set of Personal Assistant Agents (PAAs) collecting information about the context and intentions of its respective user. In each PAA were embedded four Machine Learning algorithms, with the purpose of ascertaining how wellsuited these classifiers are for filtering irrelevant POIs, in a completely automatic fashion. Supervised, incremental learning occurs when the feedback on the true relevance of each recommendation is given by the user to his PAA. To evaluate the recommendations' accuracy, we performed an experiment considering three types of users, using different contexts and intentions. As a result, all the PAA had high accuracy, revealing in specific situations F1 scores higher than 87%.

The use of machine learning algorithms in recommender systems: A systematic review

“…ML algorithms, by definition, improve their performance with access to more data. Similarly, the more data that is Baldominos et al, 2015;Banerjee et al, 2012;Bar et al, 2013;Bauer & Nanopoulos, 2014;Bellogín et al, 2011;Biancalana et al, 2011;Braida et al, 2015;Brouard & Pomot, 2016;Buabin, 2012;Cai et al, 2010Cai et al, , 2012Caraballo et al, 2014;Castro-Herrera et al, 2009;Costa et al, 2012Costa et al, , 2013Das et al, 2013;Das Dôres et al, 2016;De Gemmis et al, 2008;Diaby et al, 2013Diaby et al, , 2014Dinuzzo et al, 2011; T.-K. Fan & Chang, 2010;Forsati et al, 2009;Forsati & Meybodi, 2010;Gedikli et al, 2011;Geng et al, 2016;Haiduc et al, 2013;Hernández del Olmo et al, 2009;Hofmann, 2003Hofmann, , 2004Huang & Nikulin, 2014;Hussain et al, 2015;R. Zhang & Tran, 2010;Islam et al, 2015;Jin et al, 2005;Jun, 2005;Jung & Lee, 2004;Kao & Fahn, 2013;Karahodza & Donko, 2015;Kong et al, 2013;…”

“…Li & Chen, 2013;Tsuji et al, 2014;Xin et al, 2014;Zahra et al, 2015) Health 4 (Agarwal, 2011;Dinuzzo et al, 2011;Hussain et al, 2015;I. Song et al, 2011) Images 4 (Geng et al, 2016;Kong et al, 2013;Pantraki & Kotropoulos, 2015;Pecli et al, 2015) Tourism 4 (Costa et al, 2012(Costa et al, , 2013C. Lu et al, 2015;Sun et al, 2015) Games 3 (Baldominos et al, 2015;Castro-Herrera et al, 2009;Moreno et Another important result is the large number of studies evaluating their proposals in terms of the error in the prediction by using the RMSE (Round Mean Squared Error) and the MAE (Mean Absolute Error) metrics.…”

“…This is done for the benefit of other researchers who may be searching for an algorithm that is better than a certain threshold. (Agarwal, 2011;Aouay et al, 2014;Baldominos et al, 2015;Banerjee et al, 2012;Biancalana et al, 2011;Brouard & Pomot, 2016;Cai et al, 2010Cai et al, , 2012Caraballo et al, 2014;Costa et al, 2012Costa et al, , 2013De Gemmis et al, 2008;Devi & Venkatesh, 2013;Elmongui et al, 2015;T.-K. Fan & Chang, 2010;Z.…”

“…Song et al, 2011;Szymański & Rzeniewicz, 2016;Verma et al, 2016;Wan et al, 2009;Yap et al, 2005;Yuan et al, 2014;K. Zhao & Pan, 2015) Recall 58 32 (Aouay et al, 2014;Baldominos et al, 2015;Banerjee et al, 2012;Cai et al, 2010Cai et al, , 2012Caraballo et al, 2014;Costa et al, 2012Costa et al, , 2013De Gemmis et al, 2008;Devi & Venkatesh, 2013;Elmongui et al, 2015;T.-K. Fan & Chang, 2010;Z.…”

“…Song et al, 2011;Sun et al, 2015;Taghipour et al, 2007;Vialardi et al, 2011;Xin et al, 2014;X.-Z. Zhang, 2007) ROC 4 (Costa et al, 2012(Costa et al, , 2013X. Li & Chen, 2013;Roh et al, 2003) AUC 14 8 (Agarwal, 2011;Anaissi & Goyal, 2015;Cai et al, 2012;Diaby et al, 2013Diaby et al, , 2014…”