Mosaic benchmark network

Mosaic benchmark网络是一个用于评估动态社区检测算法的模块化链接流数据集。该数据集由里昂大学等机构创建，包含61个社区，旨在通过模拟具有预定义社区的合成模块化链接流来测试算法性能。数据集创建过程中，采用了随机种植分区网络构建连续时间边缘的方法，确保了数据集的可调基础真相。该数据集主要应用于网络科学领域，特别是在动态社区检测算法的评估和优化中，以解决在高度详细的时态网络中社区结构分析的复杂性问题。The Mosaic benchmark network is a modular link stream dataset designed for evaluating dynamic community detection algorithms. Developed by institutions including the University of Lyon, this dataset comprises 61 communities. Its core objective is to test algorithm performance by simulating synthetic modular link streams with predefined community structures. During the dataset creation process, the stochastic planted partition model was employed to construct continuous-time edges, thereby ensuring that the dataset features adjustable ground truth. This dataset is primarily utilized in the field of network science, particularly for the evaluation and optimization of dynamic community detection algorithms, aiming to resolve the complexity inherent in community structure analysis of highly detailed temporal networks.

搜集汇总

背景与挑战

背景概述

Mosaic benchmark network是一个用于评估动态社区检测算法的合成模块化链接流数据集，由里昂大学等机构创建，包含61个社区，通过随机种植分区网络方法构建，具有可调基础真相，主要应用于网络科学领域，以优化动态社区结构分析。