Towards Evaluating User Profiling Methods Based on Explicit Ratings on Item Feature

Published in Workshop on the Impact of Recommender Systems (ImpactRS) at 13th ACM Conference on Recommender Systems, 2019

Recommended citation: Yashar Deldjoo, Cristina FrĂ , Massimo Valla, Paolo Cremonesi Proceedings of the 8th Italian Information Retrieval Workshop (IIR 2017).

[Proceedings] [PDF] [bibtex]


Recommender Systems (RSs) offer a personalized support in exploring large amounts of information, assisting users in decision making about products matching their taste and preferences. Most of the research todate on recommender systems have focused on traditional users, i.e., adult individuals who are able to offer explicit feedback, write reviews or purchase items themselves. However, children’s patterns of attention and interaction are quite different from those of adults. This paper presents the rst results of a research-in-progress that can be suited to bridge the barrier between children and a recommender system by providing a child-friendly interaction paradigm. Speci cally, a web application is developed that employs real-time object recognition on movie thumbnails or DVD cover-photos in a real-time manner. The tangible object can be manipulated by the user and provide input to the system for the purpose of generating movie recommendations. We plan to extend this work to the scenario where the child could ask for a video content showing a related toy (e.g., a car, a plane, the doll of a character that she likes in a cartoon) and the system could generate the videos that matches these implicit preferences expressed by the child.