北京师范大学学报（自然科学版）2025，Vol.61Issue(6)：769-775,7.DOI:10.12202/j.0476-0301.2025143
多源数据融合的标记时间点过程方法研究
A marked temporal point process approach with multi-source data fusion
摘要
Abstract
A globally-augmented self-attentive marked temporal point process(GASAMTPP)model is proposed in this work.This model encodes an entity's static features into a global latent variable to achieve implicit clustering.The model applies a self-attention mechanism to encode historical event sequence that has been concatenated with the global latent variable to capture complex dependencies among the entity's historical events.Leveraging the resulting comprehensive representation,the conditional intensity function and the mark distribution are parameterized,enabling joint prediction of the next event time and its mark.Experiments on real datasets show that,for next-event time and mark prediction tasks,GASAMTPP outperforms existing mainstream models.关键词
标记时间点过程/自注意力机制/多源数据/隐变量Key words
marked temporal point processes/self-attention mechanism/multi-source data/latent variable分类
信息技术与安全科学引用本文复制引用
张景舒,王鹤,赵晓航..多源数据融合的标记时间点过程方法研究[J].北京师范大学学报（自然科学版）,2025,61(6):769-775,7.基金项目
国家自然科学基金资助项目(72442024) （72442024）
