A Case Elements Recognition Method Oriented to Electronic Dossier Transcript Text

Abstract

The present invention relates to the technical field of natural language processing, and relates to a case element recognition method for electronic dossier transcript text, comprising the following steps: (1) electronic dossier data preprocessing, (2) word segmentation and part-of-speech tagging in combination with a custom dictionary, (3) ) to identify the four types of case elements: time, object, weight, and amount; (4) to identify the three types of case elements: person, place, and institution. The method of the present invention fully considers the characteristics of different case elements, and adopts different processing methods in a targeted manner. For special texts such as electronic file transcripts, it can accurately identify important case elements in the absence of manual labeling data. and labeling, the case element recognition results of the present invention can be used to iteratively generate high-quality label data, thereby training a more reliable case element recognition model.

Description

technical field [0001] The invention relates to a case element recognition method for electronic file transcript text, which belongs to the technical field of natural language processing. Background technique [0002] Electronic files record and save all files generated in the process of handling cases in the form of electronic files. Electronic files have been deeply and widely used in the judicial system of our country because of their strong confidentiality, convenient statistical analysis, and high information sharing rate. With the continuous advancement of my country's "smart judiciary" informatization construction, the electronic dossier system has gradually improved, and the corresponding electronic dossier data has also increased sharply. The case handlers of the procuratorate need to review a large number of electronic files. The processing method of electronic file data is still a manual analysis and processing method. The methods and means are extremely backward,...

AI Technical Summary

Problems solved by technology

However, the current named entity recognition is mostly oriented to standardized texts such as scientific and technological documents and news reports, while electronic file transcripts have the characteristics of colloquial language expression, irregular grammar, and various sentence forms. Therefore, the existing named entity recognition methods cannot directly Applied to case element identification

Existing named entity recognition methods require a large amount of manually labeled data to train the model. Manually labeling data is an extremely laborious task, and the intelligent analysis and processing of electronic files is still in its infancy, and there is no relevant standard data set manually labeled.

At present, there is no named entity recognition method for electronic dossier transcript text, that is, there is no case element recognition method

Images