Eye-tracking and social behavior preference-based recommendation system
- 113 Downloads
With the popularization of wireless Internet technology and smartphones, the importance of recommendation systems, which analyze personality of a user using social network data such as search history, contents of written articles, the number of accesses, and etc., to achieve user convenience to obtain high profit is increasing. Since existing recommendation systems usually use only single kind of data such as social network service (SNS) data or purchase histories, the analyzed user personality by the recommendation systems can be inaccurate. Hence, in this paper, we propose an intuitive and highly accurate recommendation system by collecting personal data of a user from SNS and eye-tracking data of the user. By analyzing eye-tracking and social behaviors, we formulate preference metrics to derive category preferences. Using the preference metrics, we yield user preferences for categories. In addition, by combining and analyzing common categories between the eye-tracking and the social behaviors, we yield a final preference. Also, using the Pearson correlation coefficients, we yield the similarity between users based on the category preferences. Our experimental results show that our recommendation accuracy is 98.5% for smart TV in average and 96.5% for smartphone in average. Also, we prove that the preferences of a user can vary according to smart devices by deriving the unconscious user preferences. To derive the unconscious user preferences, we collect eye-tracking data using multiple smart devices. Consequently, the results show the applicability of our proposed scheme in a recommendation system which considers characteristics of smart devices.
KeywordsRecommendation system Eye-tracking Social media analysis Human behavior analysis
This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government MSIP) (No. 2018020767) and by the Technology development Program(C0531332) funded by the Ministry of SMEs and Startups(MSS, Korea).
- 2.John Walker S (2013) Big data: a revolution that will transform how we live, work, and think. Houghton Mifflin Harcourt, Boston, pp 181–183Google Scholar
- 4.Kim B-S, Kim H, Lee J, Lee J-H (2013) Movie recommendation system using collaborative filtering based on demographic information. In: Proceedings of KIIS Fall Conference. 23, pp 63–64Google Scholar
- 6.Chiu S-I, Hsu K-W (2017) Efficiently processing skyline query on multi-instance data. J Inf Process Syst 13:1277–1298Google Scholar
- 7.Li W et al (2014) Social recommendation algorithm dynamically adaptable to user profiling for SNS. In: 2014 Second International Conference on Advanced Cloud and Big Data (CBD), pp 261–264. IEEEGoogle Scholar
- 8.Kim J, Moon N (2017) Design of consumer behavior analysis by region through reflecting social atmosphere based on SNS. Adv Comput Sci Ubiquitous Comput. Springer, Singapore, pp 1020–1025Google Scholar
- 11.Song W, Sun G, Fong S, Cho K-E (2016) A real-time infrared LED detection method for input signal positioning of interactive media. J Converg 7:1–6Google Scholar
- 14.Chen L, Wang F (2016) An eye-tracking study: implication to implicit critiquing feedback elicitation in recommender systems. In: Proceedings of the 2016 Conference on User Modeling Adaptation and Personalization. ACM, pp 163–167Google Scholar
- 17.Kocca (2015) Content industry in 2014 statistics. Kocca. http://www.kocca.kr/cop/bbs/view/B0158897/1824539.do?searchCnd = &searchWrd = &cateTp1 = &cateTp2 = &useAt = &menuNo = 201838&categorys = 0&subcate = 0&cateCode = &type = &instNo = 0&questionTp = &uf_Setting = &recovery = &option1 = &option2 = &year = &categoryCOM062 = &categoryCOM063 = &categoryCOM208 = &categoryInst = &morePage = &pageIndex = 2. Accessed 16 Feb 2015Google Scholar