Low-frequency and easy-to-confuse criminal name prediction method integrated with case auxiliary sentences

A forecasting method and confusing technology, applied in forecasting, neural learning methods, computer components, etc., and can solve problems such as low forecasting accuracy

Active Publication Date: 2020-06-09
KUNMING UNIV OF SCI & TECH
View PDF5 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This type of method achieves good accuracy in the prediction of common crimes, but the prediction accuracy is low for low-frequency and confusing crimes. Therefore, Liu Zonglin et al. proposed a multi-task learning model for legal judgment prediction that incorporates crime keywords. The results include legal article recommendation and crime prediction. Hu et al. proposed a method that integrates crime discrimination attributes to predict low-frequency and confusing crimes. These methods are typical representatives of crime prediction research.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Low-frequency and easy-to-confuse criminal name prediction method integrated with case auxiliary sentences
  • Low-frequency and easy-to-confuse criminal name prediction method integrated with case auxiliary sentences
  • Low-frequency and easy-to-confuse criminal name prediction method integrated with case auxiliary sentences

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0052] Embodiment 1: as Figure 1-3 As shown, the low-frequency and confusing charge prediction method integrated into the case auxiliary sentence, the specific steps of the low-frequency and confusing charge prediction method integrated into the case auxiliary sentence are as follows:

[0053] Step1. Analyze the criminal case data based on the criminal case public data set, and construct the auxiliary sentence of the case; adopt a method similar to figure 1 The method analyzes a large amount of case data and constructs the case auxiliary sentences shown in Table 1. Then combine the public data sets of Chinese criminal cases to construct training sets, test sets, and verification sets;

[0054] Step2, on the basis of Step1, use the Skip-Gram model and Chars-CNN to obtain the word-level and character-level multi-granularity features of the case auxiliary sentence and case description, and then introduce a high-speed network to balance the relative contribution ratio of word ve...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a low-frequency and easy-to-confuse criminal name prediction method integrated with case auxiliary sentences, and belongs to the technical field of natural language processing. The method comprises the steps that firstly, case auxiliary sentences are constructed based on the judicial field, and the case auxiliary sentences are drafted as external knowledge to serve as mapping between case description and criminal names; calculating case description and case auxiliary sentence multi-granularity characteristics based on word level and character level; two-way mutual attention is constructed by means of case auxiliary sentences and case description, information interaction between the case description and the case auxiliary sentences is enhanced, case description multi-granularity tendency characteristics guided by the case auxiliary sentences are extracted, and accordingly low-frequency and easy-to-confuse criminal name prediction accuracy is improved. Accordingto the method, the F1 value is increased by 13.2% to the maximum extent, the accuracy is increased by 4.5% to the maximum extent, the low-frequency crime prediction F1 value is increased by 4.3%, andthe easy-to-confuse crime prediction F1 value is increased by 8.2%.

Description

technical field [0001] The invention relates to a method for predicting low-frequency and easily confusing charges integrated into case auxiliary sentences, and belongs to the technical field of natural language processing. Background technique [0002] The crime prediction task is an important subtask in the legal judgment task, which plays a vital role in the legal field. Today, the prediction accuracy of common crimes is relatively high, but the prediction accuracy of low-frequency and confusing crimes is not satisfactory, mainly because of the lack of data on low-frequency crimes and similar descriptions of confusing crimes. According to statistics, so far there are 469 categories of crimes in my country's criminal law, and the distribution of crimes is a typical long-tail distribution (a form of power-law distribution). Among the tens of millions of judgment documents data in our country, after we counted a large number of real case data, we found that relatively commo...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q10/04G06K9/62G06N3/04G06N3/08G06Q50/18
CPCG06Q10/04G06Q50/18G06N3/08G06N3/044G06N3/045G06F18/2411
Inventor 余正涛刘真丞郭军军黄于欣相艳
Owner KUNMING UNIV OF SCI & TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap