Plagiarism Detection Algorithm Based On BiLSTM And Its Application In Duplicate Checking System

With the rapid development of the Internet,network teaching mode began to rise,more and more colleges and universities have begun to gradually change from the traditional teaching mode to online teaching,many colleges and universities have launched their own online teaching system.While online teaching mode brings convenience to teaching work,it has also become a hotbed of plagiarism.Because electronic information is easy to copy and spread,the phenomenon of plagiarism is further aggravated.However,the workload of manual identification is huge,so it is particularly necessary to use computer to detect plagiarism.Based on the research of traditional plagiarism detection technology and some excellent algorithms in recent years,this paper proposes a plagiarism detection technology based on BiLSTM neural network.In view of the shortcomings of traditional plagiarism detection technology,this paper makes the following main improvements.First,in the feature extraction stage of word granularity,the pre training embedded model is used to get the word vector representation with semantic information,which solves the problem of semantic gap.Second,the BiLSTM neural network is introduced into the matching model to solve the problem that the general method can't obtain the word order information of the text.Third,after getting the feature vector of the text,the multi-layer perceptron is used to make the features of two pieces of text interact to get the text matching vector with interactive information.In this paper,the relevant data sets are used to train the model,and the experimental comparison with some commonly used methods shows that the proposed method has improved the accuracy and F1 score.Finally,the plagiarism detection module is developed based on the teaching system of the experimental center,and the model is used to detect the data of some students' homework in the school teaching system,which verifies the effectiveness of the method proposed in this paper.Finally,the design and development of plagiarism detection system are completed,and the algorithm is applied to the practice for production,which solves the problems in the actual teaching activities.