关键词: 关键词:案情要素, 行为要素, 事件类型, 时序行为链, 聚合评分器

Abstract: Abstract: Aiming at the problem that existing legal case retrieval methods lack effective utilization of case elements and were easily misled by the similarity of the semantic structure of case content, a legal case retrieval method integrating temporal behavior chain and event type was proposed. Firstly, the sequence labeling method was adopted to identify the legal event type in case description, and the temporal behavior chain was constructed by using the behavioral elements in the case text, highlighting the key elements of the case, so that the model focused on the core content of the case, so as to solve the problem that the existing methods were easily misled by the similarity of the semantic structure of the case. Secondly, the similarity vector representation matrix of the temporal behavior chain was constructed by segmented coding to enhance the semantic interaction of behavioral elements between cases. Finally, the relevance of the case was measured from three perspectives of temporal behavior chain, legal event type, and crime type through the aggregation scorer, so as to increase the rationality of the case matching score. The experimental results show that compared with the SAILER(Structure-Aware pre-traIned language model for LEgal case Retrieval) method, the P@5 score of this method is improved by 4 percentage points, the P@10 score is improved by 3 percentage points, the MAP score is improved by 4 percentage points, and the NDCG@30 score is improved by 0.8 percentage points in the LeCaRD (Legal Case Retrieval Dataset) dataset. This method effectively utilizes case elements, avoids interference from the similarity of the semantic structure of case content, and can provide a reliable basis for legal case retrieval.

Key words: Keywords: case elements, behavioral elements, event type, temporal behavioral chain, aggregation scorer