Recently, with the development of online social networks, users in social networks are usually associated with attributes such as user preferences, which is of great importance for analyzing the properties of social networks. To identify critical attributes, we propose and investigate a new problem named attribute k-core maximization. Given an attribute graph G and a budget b, we aim to identify a set of b attributes, such that the corresponding attribute k-core is maximized. Due to the NP-hardness of the problem, we resort to the greedy strategy in this paper. In order to handle large graphs, a layer-based structure and novel searching paradigms are developed to accelerate the computation. Finally, experiments over 6 real-world networks are conducted to evaluate the performance of proposed model and techniques.

近年来，随着在线社交网络的发展，社交网络中的用户通常与用户偏好等属性相关联，这对于分析社交网络的属性具有重要意义。为了识别关键属性，我们提出并研究了一个名为属性k核最大化的新问题。给定一个属性图G和一个预算b，我们的目标是确定一组b 个属性，使得相应的属性k-core 最大化。由于问题的 NP 难度，我们在本文中采用贪心策略。为了处理大图，开发了基于层的结构和新颖的搜索范式来加速计算。最后，进行了 6 个真实世界网络的实验，以评估所提出模型和技术的性能。