SVM significant role selection method for improving semantic text plagiarism detection

Abstract

This research introduces an approach for the prediction and detection of plagiarized text based on Semantic Role Labelling (SRL) and Support Vector Machine (SVM). The introduced method evaluates and analyses text based on semantic position for each term within the text. It additionally detects the source semantic sense in considering the connections between its terms using the Semantic Role Labeling (SRL). SRL presents noteworthy remuneration while creating roles from a text semantically. Selecting for every role created by the SVM method keeping in mind the end goal to foresee significant roles is a noteworthy part of the proposed system. The imperative roles that will vote by the SVM strategy will be chosen in the comparability computation process. The proposed strategy assessed utilizing the PAN-PC-10 dataset. The outcomes proved that the introduced strategy enhanced the execution as far as the assessment measures contrasted and other plagiarism detection methods.