Automatic construction method for legal instrument and action cause classifier

Abstract

The invention provides an automatic construction method for a legal instrument and action cause classifier and relates to the field of knowledge engineering technology. The classifier is high in classification efficiency and high in accuracy. The method comprises the steps that legal instruments of multiple types of action causes are acquired, and a test corpus of each type of action causes is determined; stop words and segmented words are removed from each type of test corpus; a basic keyword of each type of action causes is determined according to a word segmentation result; word vector training is performed on each type of test corpus according to the word segmentation result, the keywords are expanded, and an expanded keyword list of each type of action causes is determined according to the expanded keywords and the determined basic keywords; a to-be-classified legal instrument is input, stop words and segmented words are removed from the input to-be-classified legal instrument, and word vector training is performed on the input to-be-classified legal instrument; and a semantic similarity sum of the to-be-classified legal instrument and each type of expanded keyword list is calculated according to a word vector training result of the to-be-classified legal instrument, and the action cause corresponding to the maximum semantic similarity sum is taken to serve as the action cause corresponding to the to-be-classified legal instrument.

Description

technical field [0001] The invention relates to the technical field of knowledge engineering, in particular to an automatic construction method of a classifier for legal documents. Background technique [0002] In recent years, in order to implement the principle of open trial, the Supreme People's Court of our country has established the China Judgment Documents Network, stipulating that it is not appropriate to publish on the Internet except those involving state secrets, personal privacy, juvenile crimes, cases closed through mediation, and other cases. Except for the judgment documents, the rest will be published on the Internet. This policy brings a large amount of research resources to natural language processing technology for legal documents. [0003] In the context of the era of legal information disclosure, in the existing massive legal documents, let the computer acquire learning ability and more accurate domain knowledge, and effectively use these massive text r...

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is to provide a method for automatically constructing a case classifier for legal documents, so as to solve the problems of low efficiency and low accuracy of case extraction existing in the prior art

Images