Plagiarism Awareness among Research Scholars: Wasserstein Deep Convolutional Generative Adversarial Network Approach for Detection in Tamil Nadu Universities

Abstract

Detecting paraphrases in Indian languages requires a thorough examination of lexical, syntactic, and semantic characteristics. Because of structural variations from languages such as English, the use of lexico-syntactic characteristics differs across Indian languages, which has a substantial impact on system performance. In this manuscript, Plagiarism Awareness among Research Scholars: Wasserstein Deep Convolutional Generative Adversarial Network Approach for Detection in Tamil Nadu Universities (PA-WDCGAN-TNU) is proposed. Initially the data is collected from DPIL@FIRE 2016 shared task dataset for Tamil Indian language. Then, data are fed to pre-processing segment. For pre-processing Probability-Based Synthetic Minority Oversampling Technique (P-SMOT) is used for tokenization, stemming and removal of stop words. Then, preprocessed data is given into Wasserstein Deep Convolutional Generative Adversarial Networks (WDCGAN) to evaluate paraphrase sentences in Tamil language and categorize the classes like “Paraphrase” and “Non-Paraphrase”. In Generally, WDCGAN doesn’t expose some adoption of optimization systems for calculating optimal parameters for exact classifies lung cancer. Hence, Seasons Optimization Algorithm (SOA) is proposed to enhance WDCGAN that exactly classify paraphrase sentences in Tamil language. The proposed PA-WDCGAN-TNU technique is executed in Python and performance metrics likes accuracy, precision, recall, FI-score, computational times are evaluated. The simulation results display that Performance of the PA-WDCGAN-TNU approach attains 29.8%, 21.2%, and 18.9% higher Accuracy, 24.7%, 32.5%, and 29.6% higher precision, 25.8%, 28.5%, and 21.6% higher Recall are analyzed with existing methods like Tamil Paraphrase Detection Utilizing Encoder-Decoder Neural Networks (TPD-EDNN), Novel Technique for Developing Paraphrase Detection System utilizing Machine Learning (PD-ML)and Paraphrase Detection in Indian Languages Utilizing Deep Learning (PD-IL-DL)  respectively..