A versatile package recommendation framework aiming at preference score maximization
- PDF / 2,366,347 Bytes
- 19 Pages / 595.276 x 790.866 pts Page_size
- 111 Downloads / 198 Views
ORIGINAL PAPER
A versatile package recommendation framework aiming at preference score maximization Panagiotis Kouris1 · Iraklis Varlamis2 · Georgios Alexandridis1 · Andreas Stafylopatis1 Received: 5 January 2018 / Accepted: 13 May 2018 © Springer-Verlag GmbH Germany, part of Springer Nature 2018
Abstract Package recommendation systems have gained in popularity especially in the tourism domain, where they propose combinations of different types of attractions that can be visited by someone during a city tour. These systems can also be applied in suggesting home entertainment, proper nutrition or academic courses. Such systems must optimize multiple user criteria in tandem, such as preference score, package cost or duration. This work proposes a flexible framework for recommending packages that best fit users’ preferences while satisfying several constraints on the set of the valid packages. This is achieved by modeling the relation between the items and the categories these items belong to, aiming at recommending to each user the top-k packages that cover their preferred categories and the restriction of a maximum package cost. Our contribution includes an optimal and a greedy algorithm, that both outperform a state-of-the-art system and a popularity-based baseline solution. The novelty of the optimal algorithm is that it combines the collaborative filtering predictions with a graph-based model to produce package recommendations. The problem is expressed through a minimum cost flow network and is solved by integer linear programming. The greedy algorithm has a low computational complexity and provides recommendations which are close to the optimal one. An extensive evaluation of the proposed framework has been carried out on six popular recommendation datasets. The results obtained using a set of widely accepted metrics show promising performance. Finally, the formulation of the problem for specific domains has also been addressed. Keywords Recommendation system · Package recommendations · Top-k packages · Collaborative filtering
1 Introduction Recommender systems (RSs) have become popular since they can personalize a user’s experience by providing automated recommendations. RSs operate by analyzing user * Panagiotis Kouris [email protected] Iraklis Varlamis [email protected] Georgios Alexandridis [email protected] Andreas Stafylopatis [email protected] 1
School of Electrical and Computer Engineering, National Technical University of Athens, 9, Iroon Polytechniou str, Zografou, 15780 Athens, Greece
Department of Informatics and Telematics, Harokopio University of Athens, 9 Omirou str, Tavros, 17778 Athens, Greece
2
preference data, trying to identify correlations between them. User preference is expressed in various forms, such as the history of purchases, usage logs and numerical ratings in a predefined scale (e.g. five-star rating system). The RSs, in return, may propose a variety of items; for example, an online shop could suggest books or movies to users based on their profile, thei
Data Loading...
