Online Estimation of Multiple Dynamic Graphs in Pattern Sequences

The dataset used in the paper is a binary time-series data, where each time step is a binary pattern. The data is generated from a mixture of multiple graphs with time-dependent weights.

本论文采用的数据集为二元时间序列数据（binary time-series data），其中每个时间步均对应一个二元模式。该数据集由权重随时间动态变化的多图混合体生成。