Platform-Oriented Event Time Allocation
Heli Sun,Ning Wang,Jingyu Jia,Jianbin Huang,Hui Xiong,Liang He,Xinwang Liu,Shan Zhang,Shaojie Qiao,Jizhong Zhao
DOI: https://doi.org/10.1109/tkde.2021.3109838
IF: 9.235
2021-01-01
IEEE Transactions on Knowledge and Data Engineering
Abstract:Online Event-based social networks (EBSNs), such as Meetup and Whova, which provide platforms for users to publish, arrange and participate in events, have become increasingly popular. A major challenge for managing EBSNs is to generate the most satisfactory event arrangement, i.e., events are scheduled at the reasonable time to attract maximum number of participants. Existing approaches usually focus on assigning a set of events organized to time intervals, but ignore the competitive relationships among different event organizers, which will lead to event time allocations unacceptable to organizers. Thus, a more intelligent EBSNs platform that allocates social events properly in a global view (i.e., the perspective of platform) is desired. In this paper, we first formally define the problem of Platform-oriented Event Time Allocation (PETA), which contains two parts: the prediction of event feasible time period and the event time allocation. Unfortunately, we find that the PETA problem is NP-hard due to the global conflict constraints on events. Thus, we propose a method to calculate event feasible time period based on event time prediction, and then design a greedy algorithm and two approximation algorithms to solve the PETA problem. Finally, we conduct extensive experiments on both real and synthetic datasets to test the effectiveness and efficiency of the proposed algorithms.
computer science, information systems, artificial intelligence,engineering, electrical & electronic