Method and device for labeling referee documents based on machine learning algorithm

Abstract

The invention discloses a method and device for marking referee documents based on a machine learning algorithm. The method includes: collecting a text set of referee documents to be marked; performing structural segmentation on the text in the text set; establishing a semantic tag library; manually marking the referee documents to be marked based on the semantic tag library; The marked referee documents are used as a standard data set to be handed over to machine learning to train and optimize the preliminary labeling model; the remaining part of the manually marked referee document samples are selected as the verification data set, which is used to improve the semantic label library, and the preliminary labeling The model is iterated and optimized to obtain a mature referee document labeling model; the target referee document to be marked is structurally segmented and then input into the mature referee document labeling model to obtain the labeling result. The invention solves the problems of incomplete extraction of legal elements of judgment documents and low accuracy of extraction of case information in the related art.

Description

technical field [0001] The present invention relates to the technical field of text processing, in particular to a method and device for labeling referee documents based on machine learning algorithms. Background technique [0002] The content of judgment documents is usually relatively long, and the facts of some cases are more complicated. For judges, it is a great challenge to quickly capture the key information points of the case and clarify the judicial logic of the case from similar cases that are pushed, which is time-consuming and expensive. It is laborious and puts pressure on the judge's trial work. For adjudication documents, it is often necessary to analyze the type of adjudication documents, each paragraph, and extract the dimension information of each paragraph and other parameters, so as to summarize historical cases in a timely manner and discover laws to improve trial efficiency and achieve judicial justice. At the same time, analyzing the effective judgmen...

AI Technical Summary

Problems solved by technology

This method only considers the method of using keywords to extract legal elements, but its shortcomings are very prominent: the judgment document is not structurally divided, and the repetition of keywords leads to a decrease in the accuracy of case element extraction; the general use of case keywords to replace the concept of legal elements, Lack of professionalism in extracting legal elements of judgment documents, which does not conform to realistic judicial logic

[0008] Therefore, those skilled in the art are committed to developing a method and device for labeling referee documents based on machine learning algorithms. Through the type labeling method based on natural language understanding and semantic analysis, the referee documents are marked and expressed in an intuitive way. The complete legal elements and internal logical relationship of the case are solved, thus solving the problems of incomplete extraction of legal elements of judgment documents and low accuracy of case information extraction caused by ignoring legal expertise and judicial logic in existing methods

Images