Neural Hawkes Process Datasets
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/HMEIatJHU/neural-hawkes-particle-smoothing
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了由不同参数的神经霍克斯过程生成的事件序列，用于评估所提出粒子平滑方法的性能。数据集涵盖了确定性缺失机制与随机性缺失机制。在模型训练和提案分布过程中，使用了完全观察到的数据。数据集的规模从平均长度为15的短序列到平均长度超过300的长序列不等。该任务旨在对连续时间事件流中的缺失事件进行插补。
This dataset contains event sequences generated by Neural Hawkes Processes with different parameters, which is used to evaluate the performance of the proposed particle smoothing method. The dataset covers both deterministic and stochastic missingness mechanisms. Fully observed data is employed during model training and proposal distribution derivation. The sequence lengths range from short ones with an average length of 15 to long ones with an average length exceeding 300. This task aims to impute missing events in continuous-time event streams.
提供机构：
Synthetic generation based on neural Hawkes processes
