Beyond User Preferences: The Short-Term Behaviour Modelling Michal Kompan Ondrej Kassak Maria Bielikova Slovak University of Technology in Slovak University of Technology in Slovak University of Technology in Bratislava, Faculty of Informatics and Bratislava, Faculty of Informatics and Bratislava, Faculty of Informatics and Information Technologies Information Technologies Information Technologies Ilkovicova 2 Ilkovicova 2 Ilkovicova 2 Bratislava, Slovak Republic 842 16 Bratislava, Slovak Republic 842 16 Bratislava, Slovak Republic 842 16 name.surname@stuba.sk name.surname@stuba.sk name.surname@stuba.sk ABSTRACT As the researchers attention is usually addressed to user prefer- The context of a user is a notoriously researched topic in the recom- ences, user behavior (or the behavior change) is not considered in mender systems community. It greatly influences user preferences recommender approaches. We understand the behavior as a set of and respectively his/her behaviour. The research focuses on the actions, a user performs on the web. In other words, while prefer- actual influence affecting user and temporal preferences of users. ences express what the user likes, the behavior describes how the These tell us what the user likes, but fail at describing his/her be- user acts. Similarly to preferences [5, 14], we distinguish between havior. We believe that the user’s actual behavior represents a great short- (within a session) and long-term behavior. source of information, which is useful for recommendation meth- We believe, that considering a user short-term behavior in mod- ods. By modelling the user short-term behavior (on the level of a ern recommendation approaches, will improve the user experience. session), we are able to, e.g., predict the session end intent or the Moreover, a prediction of such behavior opens new opportunities length of the session and respectively adjust generated recommen- for e-commerce to adjust recommendations and to optimize the dations. As a result, users benefit from a seamless user experience utility functions (e.g., profit). In this paper, we provide a brief dis- and the business utilizes its objective functions (e.g., profit). cussion on pros and cons of the short-term user behavior modelling and its application to the personalized recommendation problem. CCS CONCEPTS • Information systems → World Wide Web; Personalization; 2 USER PREFERENCES VS. BEHAVIOR Web mining; Personalization; The major of a research activity is focused on exploring acquisi- tion [2], modeling [11] and further usage [8] of user preferences. KEYWORDS These are the base for almost all modern Web-based services. user context, short-term behavior, personalization, user model As the next step, a user context proved to be extremely helpful in the personalization [15]. As the preferences of the user are rich, ACM Reference format: the context reduces a set of preferences or items, which have to be Michal Kompan, Ondrej Kassak, and Maria Bielikova. 2017. Beyond User considered by the recommender approach. Both these concepts tell Preferences: The Short-Term Behaviour Modelling. In Proceedings of Tem- us what the user likes. poral Reasoning in Recommender Systems Workshop @ ACM RecSys, Como, Italy, August 2017 (TRRS), 3 pages. On the contrary, a user behavior describes how a user acts. More specifically, the short-term behavior describes user actions within a session (in the Web context). This behavior is based on a user inter- 1 INTRODUCTION action with the content and the site structure. For instance, what Personalized recommendation became an integral part of modern time the user spent in actual session, how many items he/she vis- Web applications. User preferences serve as the basis for tailoring ited, how many hyperlinks there are available to follow. It has been services and its content for specific user. This was proved to work proven that the website usage information improves the personal- very well in several domains [3]. Despite of this, researchers found ization performance [13]. We believe that the knowledge of user out that user preferences are highly influenced by his/her context, typical and/or future behavior allows to improve recommendations e.g., time, location, mood. Context is generally defined as "a prede- – bringing benefits both for users and commerce as well. fined set of contextual attributes, the structure of which does not Surprisingly, despite its potential, the user short-term behavior change over time" [1]. is not researched from the personalization point of view. It is clear The context helps us to identify an actual user state, which that the behavior is also influenced by a user actual state and thus reflects to some extent the user’s objectives. Usually, long-term the context has to be somehow taken into account. Nevertheless, it preferences are modelled by user models. A user model is defined offers plenty of research questions and challenges. as a set of information connected with the user behaviour, attitudes and stereotypes [4]. Together, both short-term and long-term user 3 SHORT-TERM BEHAVIOR MODELLING preferences help to generate relevant recommendations. From the behavior modelling perspective, the short-term behavior (within a session) brings several opportunities for the recommen- Copyright ©2017 for this paper by its authors. Copying permitted for private and academic purposes. dation process. The short-term behavior is similarly to user prefer- ences, influenced by his/her context. From this point of view, the TRRS, August 2017, Como, Italy M. Kompan et al. site itself (e.g., content, structure) represents a major contextual information. In the literature three aspects are usually reported as helpful to model the user behavior on a site [10]: website structure; website content and website usage. Despite of the importance of the short-term behavior, it is al- ways based on the long-term behavior. The short-term behavior itself describes actual user actions, but it is very noisy. There is a great chance that the user actual behavior will follow in some measure his/her historic patterns. Moreover, we act as individuals, but generally we are subjects of some actual trends (e.g., following breaking news, learning session before exams) [7]. For this reason, Figure 1: The idea of short-term behavior user model [9]. user actual behavior may by similar to behavior of other users. To reflect all these characteristics, in our previous work, we proposed a domain independent short-term user model [9]. Follow- session and a user who will perform 20 actions within a session. For ing the idea of the short-term and long-term behavior, our model instance, an educational system should recommend an important consists (Fig. 1) of several time layers (e.g., day, week), two parts learning object for a student, who is about to leave the system. (personal – reflecting a behavior of modelled user; global – reflect- On the contrary, more exercises can be recommended to a student ing a behavior of all users). Moreover, several additional attributes performing a long session. Such an approach is clearly beneficial for are stored to help describe the (temporal) actual session (e.g., time the students by trying to maximize their knowledge in the specific spent within the session). amount of time. As a result, we monitor several comparative characteristics for Similarly, based on the session end intent prediction, a content every time layer (e.g., hour, day, week) and part (comparison to user provider can decide which personalized adverts (considering user previous behavior and also to all users). As the model considers context) should be displayed to the customer. Another example of only characteristics (e.g., number of sentences, images), model is application such a concept is to offer a discount or special offer. domain independent. Based on our experience, with the short-term behavior modelling, Afterwards the short-term behavior of the user is modelled, we we identified following challenges: often want to predict the user future behavior. From the website Short-term behavior modelling. User behavior is a valuable point of view, the user session end intent is an important task. In the source of information for modern web-based services. We e-commerce this is observed from the long-term perspective (e.g., believe, that similarly to the user preferences also a user contract renewal) and referred as the user attrition [12]. On the Web, behavior is influenced by his/her context. Considering the this is quite a novel task, which needs to be further explored [6]. user context can further improve the short-term modelling In our previous works, we made first steps to explore the user and its consequent applications. session end intent prediction [7]. We used a binary linear classifier Session end intent prediction. As our experiments showed, over a stream of actions from news and e-learning domains. As our we are able to predict the user short-term behavior on a web- results indicate, thanks to proposed model, we are able to predict the site. A more sophisticated machine learning approach may user end intent within three actions in advance (precision 82%) [9]. further improve the performance. On the other hand, one This is a promising result, which indicates that proposed model of the implicit requirements is the prediction in online time adapts various domain characteristics and reflects slight changes (as we need to react to the actual user session). These two of the user behavior. aspects need to be balanced. Hand by hand with the session end intent prediction, the time Spend time prediction. In some domains, also a time, the spent prediction on the site seems to be a promising task. In some user will spend by browsing, represents a valuable informa- domains, not only number of actions, but also an approximated time, tion. Two basic approaches have to be explored – a classi- the user will browse may bring better user experience and business fication to predefined classes (e.g., time intervals) and a re- potential. All of these tasks use the user short-term behavior and gression in order to approximate specific time. we believe that will improve the quality of web services. Short-term behavior boosted recommendation. The most interesting application of the short-term behavior is its con- sidering in recommendation. There are several options as 4 CONCLUSIONS to selecting different recommender method or to prioritize There are no doubts that the user short-term behavior is an im- specific items. This application is mostly appreciated by the portant source of information and knowledge, describing the user. commerce. Together with the user context and his/her preferences allow mod- ern Web-based services react to actual user surroundings and tastes. ACKNOWLEDGMENTS The short-term behavior modelling [9] aims at capturing slight This work was partially supported by grant No. APVV-15-0508: behavior changes by comparing actual to historical user behavior Human Information Behavior in the Digital Space and by grant and also to the behavior of the rest of users. The knowledge of No. VG 1/0646/15: Adaptation of access to information and knowl- user future actions allows us to differentiate recommendations, edge artifacts based on interaction and collaboration within web e.g., for a user who will perform two actions before leaving the environment. Beyond User Preferences: The Short-Term Behaviour Modelling TRRS, August 2017, Como, Italy REFERENCES [1] Gediminas Adomavicius and Alexander Tuzhilin. 2011. Context-Aware Recom- mender Systems. Springer US, Boston, MA, 217–253. https://doi.org/10.1007/ 978-0-387-85820-3_7 [2] Joshua Akehurst, Irena Koprinska, Kalina Yacef, Luiz Pizzato, Judy Kay, and Tomasz Rej. 2012. Explicit and Implicit User Preferences in Online Dating. Springer Berlin Heidelberg, Berlin, Heidelberg, 15–27. https://doi.org/10.1007/ 978-3-642-28320-8_2 [3] Shlomo Berkovsky and Jill Freyne. 2015. Web Personalization and Recommender Systems. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD ’15). ACM, New York, NY, USA, 2307–2308. https://doi.org/10.1145/2783258.2789995 [4] Peter Brusilovsky. 1996. Methods and techniques of adaptive hypermedia. User Modeling and User-Adapted Interaction 6, 2 (1996), 87–129. https://doi.org/10. 1007/BF00143964 [5] Dietmar Jannach, Lukas Lerche, and Michael Jugovac. 2015. Adaptation and Evaluation of Recommendations for Short-term Shopping Goals. In Proceedings of the 9th ACM Conference on Recommender Systems (RecSys ’15). ACM, New York, NY, USA, 211–218. https://doi.org/10.1145/2792838.2800176 [6] Ondrej Kassak, Michal Kompan, and Maria Bielikova. 2015. Web User Short- term Behavior Prediction: The Attrition Rate in User Sessions. In Proceedings of the ECMLPKDD 2015 Doctoral Consortium (Aalto University publication series Science+Technology), Jaakko HollmÃľn and Panagiotis Papapetrou (Eds.). Aalto University; Aalto-yliopisto, 272. http://urn.fi/URN:ISBN:978-952-60-6443-7 [7] Ondrej Kassak, Michal Kompan, and Maria Bielikova. 2016. Student behavior in a web-based educational system: Exit intent prediction. Engineering Applications of Artificial Intelligence 51 (2016), 136 – 149. https://doi.org/10.1016/j.engappai. 2016.01.018 Mining the Humanities: Technologies and Applications. [8] Ondrej Kassak, Michal Kompan, and Maria BielikovaËĞ. 2016. Personalized hybrid recommendation for group of users: Top-N multimedia recommender. Information Processing & Management 52, 3 (2016), 459 – 477. https://doi.org/10. 1016/j.ipm.2015.10.001 [9] Michal Kompan, Ondrej Kassak, and Maria Bielikova. 2017. The Prediction of Session End Intent: The Short-term User Model. (2017). World Wide Web, submitted. [10] Raymond Kosala and Hendrik Blockeel. 2000. Web Mining Research: A Survey. SIGKDD Explor. Newsl. 2, 1 (June 2000), 1–15. https://doi.org/10.1145/360402. 360406 [11] Tomáš Kramár, Michal Barla, and Mária Bieliková. 2013. Personalizing Search Using Socially Enhanced Interest Model, Built from the Stream of User’s Activity. J. Web Eng. 12, 1-2 (Feb. 2013), 65–92. http://dl.acm.org/citation.cfm?id=2481562. 2481565 [12] Fan Li, Juan Lei, Ying Tian, Sakuna Punyapatthanakul, and Yanbo J. Wang. 2011. Model Selection Strategy for Customer Attrition Risk Prediction in Retail Banking. In Proceedings of the Ninth Australasian Data Mining Conference - Volume 121 (AusDM ’11). Australian Computer Society, Inc., Darlinghurst, Australia, Australia, 119–124. http://dl.acm.org/citation.cfm?id=2483628.2483643 [13] Bamshad Mobasher, Robert Cooley, and Jaideep Srivastava. 2000. Automatic Personalization Based on Web Usage Mining. Commun. ACM 43, 8 (Aug. 2000), 142–151. https://doi.org/10.1145/345124.345169 [14] Wei Wang, Dongyan Zhao, Haining Luo, and Xin Wang. 2013. Mining User Interests in Web Logs of an Online News Service Based on Memory Model. In Proceedings of the 2013 IEEE Eighth International Conference on Networking, Architecture and Storage (NAS ’13). IEEE Computer Society, Washington, DC, USA, 151–155. https://doi.org/10.1109/NAS.2013.25 [15] Yong Zheng, Bamshad Mobasher, and Robin Burke. 2015. Similarity-Based Context-Aware Recommendation. In Proceedings, Part I, of the 16th Interna- tional Conference on Web Information Systems Engineering — WISE 2015 - Vol- ume 9418. Springer-Verlag New York, Inc., New York, NY, USA, 431–447. https: //doi.org/10.1007/978-3-319-26190-4_29