关键词: 深度聚类, 文本聚类, 事件表示, 事件结构, 对比学习

Abstract: Aiming at the problem that the existing deep clustering methods is difficult to efficiently divide event types without considering event information and its structural characteristics, a Deep Event Clustering method based on Event Representation and Contrastive Learning (DEC_ERCL) was proposed. Firstly, information recognition was utilized by this method to identify structured event information from unstructured text, thus the impact of redundant information on event semantics was avoided. Secondly, the structural information of the event was integrated into the autoencoder to learn the low-dimensional dense event representation, which was used as the basis for downstream clustering. Finally, in order to effectively model the subtle differences between events, a contrast loss with multiple positive examples was added to the feature learning process. Experiments results on the datasets DuEE, FewFC, Military(a dataset of domain events based on open-source Military news) and ACE2005 show that the proposed method performs better than other deep clustering methods in accuracy and normalized mutual information evaluation indexes. Compared with the suboptimal IDEC(Improved Deep Embedding Cluster), the accuracy of DEC_ERCL is increased by 21.34%, 26.46%, 7.36% and 39.97%, respectively, indicating that DEC_ERCL has better event clustering effect.

Key words: deep clustering, text clustering, event representation, event structure, contrastive learning