Recomendação de Conteúdo e QoE: Um Experimento Quantificando o Papel da QoS nas Preferências por Vídeos
Abstract
Recommender systems are increasingly present in Internet users’ routine. Therefore, platforms like Youtube and Netflix seek to improve their recommendation systems, to provide a better experience for their users. However, the users’ experience depends on a multitude of factors. In particular, caching systems have an important influence in the quality of experience (QoE), since they impact quality of service (QoS) metrics (such as the delay and the throughput) experienced by users. Our goal is to study the viability of a QoS-aware and QoE-friendly content recommendation system. To this aim, we conduct an experiment with real users, having different profiles. Each user is requested to evaluate different videos, which vary in their contents and in the corresponding QoS. Given our findings on QoS-QoR tradeoffs, we investigate their impact on the design of a recommendation system. A decision tree classifier reached accuracy of 77% using cross validation, which allows us to further understand the user’s decision making process.
References
de Freitas, R. G., Menasché, D., Delgado, C., and Ziviani, A. (2017). Recomendação de conteúdo e desempenho de sistemas de cache.
Dehghan, M., Massoulie, L., Towsley, D., Menasche, D., and Tay, Y. C. (2016). A utility optimization approach to network cache design. In INFOCOM2016, pages 1–9. IEEE. Gomez-Uribe, C. A. and Hunt, N. (2016). The netflix recommender system: Algorithms, business value, and innovation. ACM Transactions on Management Information Sys- tems (TMIS), 6(4):13.
Hutto, C. J. and Gilbert, E. (2014). Vader: A parsimonious rule-based model for sentiment analysis of social media text. In AAAI. Krishnan, S. S. and Sitaraman, R. K. (2013). Video stream quality impacts viewer beha- vior: inferring causality using quasi-experimental designs. IEEE/ACM Transactions on Networking, 21(6):2001–2014.
Liu, D. and Yang, C. (2018). A learning-based approach to joint content caching and recommendation at base stations. arXiv preprint arXiv:1802.01414.
Mok, R. K., Chan, E.W., and Chang, R. K. (2011). Measuring the quality of experience of http video streaming. In IFIP/IEEE Integrated Network Management, pages 485–492.
Nam, H., Kim, K.-H., and Schulzrinne, H. (2016). Qoe matters more than qos: Why people stop watching cat videos. In INFOCOM 2016, pages 1–9. IEEE.
Sammut, C. andWebb, G. I. (2017). Encyclopedia ofMachine Learning and Data Mining. Springer Publishing Company, Incorporated, 2nd edition. Sermpezis, P., Spyropoulos, T., Vigneri, L., and Giannakas, T. (2017). Femto-caching with soft cache hits. In GLOBECOM, pages 1–7. IEEE.
Turnbull, D. (2017). High-quality recommendation systems with elasticsearch. https: //tinyurl.com/elasticc.
Véras, D., Prudêncio, R., Ferraz, C., Bispo, A., and Prota, T. (2015). Context-aware techniques for cross-domain recommender systems. In 2015 Brazilian Conference on Intelligent Systems (BRACIS), pages 282–287. IEEE.
