Research On Contract Risk Detection Algorithm And System Implementation

In recent years,the protection of copyright has gradually become a hot issue in the society.Due to the increasingly complex network environment,personal works are spread through multiple channels on the Internet.In most cases,the channels,methods and objects of transmission are lack of strong protection.Copyright infringement cases are common.Many copyright owners are actively seeking effective ways to protect their legitimate rights and interests of copyright.The signing of copyright license contract can provide protection for copyright owners to a great extent,but only when the contract terms are in line with the relevant laws and regulations,and the contract items are perfect and accurate,can it play its protective role to the greatest extent.A risk contract is likely to bring unpredictable losses to both parties.Although the traditional way of manual evaluation and detection of contracts can greatly avoid risks,it will bring a lot of economic costs and time costs,so people need a more intelligent and efficient detection method.Based on the above background and problems,this paper proposes a contract text risk detection algorithm for the field of copyright license.Based on the relevant laws and regulations of the contract for the use of copyright license,this paper summarizes the standard elements of copyright contract,matches the contract text with the contract standard elements,and tests whether the contract text contains all the standard elements of copyright license use,such as if one element is not If the contract text is not standardized,it is considered as a risk element,and a risk detection result is given.Based on this algorithm,the contract risk detection system for copyright license is designed and implemented.Compared with the traditional manual inspection method,the risk detection of contract text directly through contract risk detection system can minimize the time and economic cost of the detection.The contract elements summarized by the laws and regulations on the use of copyright license contract can also guarantee the professionalism of the test results,and provide users with more convenient and reliable risk detection services.The main contributions of this paper include:Firstly,this paper studies and summarizes the relevant laws and regulations of copyright license contract,such as copyright law,regulations on the implementation of copyright law and civil code,and summarizes the standard elements of the contract for the use of copyright,including 9 first-class elements,14 secondary elements and corresponding descriptions of these secondary elements.The essential elements database provides a reliable basis for contract risk detection.Secondly,in this paper,a comprehensive text feature extraction method based on text similarity is proposed,which measures the similarity between texts from multiple dimensions such as morphology,part of speech,syntax and semantics,including feature extraction based on BERT,feature extraction based on dependency parsing and part of speech tagging,feature extraction based on literal matching similarity and feature extraction based on semantic role tagging similarity Feature extraction consists of four parts.Thirdly,based on the constructed standard elements Library of copyright licensing contract and the comprehensive text feature extraction method,this paper proposes a contract risk detection algorithm.Firstly,the contract text is matched with the contract standard elements library,and the feature vector of the contract text is obtained through the feature extraction of the matching result by the comprehensive text feature extraction method;then the risk classification prediction of each element is carried out by using the contract risk detection model,and the detection result of 0 indicates that the element is a risk element,and the detection result of 1 indicates that the element is a safety element.In the model construction stage,this paper selects the deep neural network model,selects the best risk detection model through training and optimization,and returns the contract risk detection results for users.