ساخت گراف تکاملی برای توصیه متوالی
Abstract
1. Introduction
2. Related work
3. Evolving graph construction for event recommendation
4. Experiment results
5. Conclusion
Appendix.
References
Abstract
Personalized recommendation can help individual users to quickly reserve their interested events, which makes it indispensable in event-based social networks (EBSNs). However, as each EBSN is often with large amount of entities and each upcoming event is normally with non-repetitive uniqueness, how to deal with such challenges is crucial to the success of event recommendation. In this paper, we propose an evolving graphbased successive recommendation (EGSR) algorithm to address such challenges: The basic idea is to exploit the random walk with restart (RWR) on a recommendation graph for ranking the upcoming events. In EGSR, we employ a sliding window mechanism to construct evolving graphs for successively recommending new events for each user. We propose a graph entropy-based contribution measure for adjusting the window length and for weighting the history information. In EGSR, we also apply a topic analysis technique for analyzing event text description. We then propose to establish each user an interest model and to compute the similarities in between event content and user interest as edges’ weights for each recommendation graph. In successive recommendation, the number of upcoming events may experience great variations in different times. For a fair comparison, we also propose a set of cumulative evaluation metrics based on the traditional recommendation performance metrics. Experiments have been conducted based on the crawled one year data from a real EBSN for two cities. Results have validated the superiority of the proposed EGSR algorithm over the peer ones in terms of better recommendation performance and reduced computation complexity.
Introduction
With the fast development of Internet of Things (IoT), recent years have witnessed the emergence of a new computing paradigm, called Cybermatics, which have been continuously integrating diverse Cyber, Physical, and Social Systems and promoting numerous new applications everyday [1]–[۴]. For example, given the wide adoption of smartphones, people can arrange their daily life more convenient and expand their social circles [5]–[۷]; While with the population of eventbased social networks (EBSNs), people can easily reserve their interested events through their smartphones [8]–[۱۰]. However, due to the proliferation of online events, how to accurately recommend individual users their mostly interested ones becomes a challenging task. Although some EBSNs, such as Meetup and Douban Event 1 , provide a search function for users to find their preferred events with key words, how to accurately match user preferences with appropriate events is still very difficult, especially for most users being unable to clearly express their interests. In response to the pressing demands, a good event recommendation system is much required for EBSNs. Event recommendation in EBSNs often faces the cold start problem [11], [12]. Compared with the general item recommendation, like recommending books and movies, events are usually with the property of non-repetitive uniqueness [13]. Furthermore, an upcoming event generally cannot be actually ’consumed’ and evaluated, though it may be reserved by some users, before its commencement. To deal with such challenges, we can exploit the history events that a user had once attended to establish an interest model for him. Among many event properties, like the launching time and place, we believe that the event text description could provide more intrinsical information for reflecting users’ interests. So it is necessary to analyze event text description, which can be done by enjoying some recently developed topical analysis techniques [14], [15].