<![CDATA[Plagiarism is an act of imitating, quoting and even copying or acknowledging the work of others as one’s own work. A final project is one of the mandatory requirements for students to complete learning at college. It must be written by the students based on their own ideas. However, there is much plagiarism because it is easy to carry out just by simply copying the text of other people’s ideas and then pasting it into a worksheet and admitting that the ideas are theirs. In addition, replacing some words in other people’s sentences with their own language style without properly acknowledging the original source of the quotation is also an act of plagiarism. A manual check for the final project also becomes an issue for the final project coordinator, i.e., it needs high accuracy and a relatively long time to check the plagiarism in the final project document. Therefore, implementing plagiarism detection mechanisms is necessary to mitigate the escalation of plagiarism occurrences. In response to those matters, this study aims to design a system capable of identifying textual similarities by focusing on sentences containing synonymous words. One of the used algorithms is synonym recognition, which detects words that possess synonymous meanings by comparing each term with the entries in a dictionary. The synonym recognition is combined with the winnowing method, functioning as a fingerprint-based text weighting. After the weight of each document is obtained, the similarity level between documents is calculated with the cosine similarity algorithm. The inclusion of synonym recognition in conjunction with the winnowing weighting method resulted in a notable gain of 3.11% in the average similarity scores for title and abstract detection, compared to the absence of synonym recognition. The results show that the used algorithms are accurate with accuracy testing and root mean squared error (RMSE).]]>
Copyrights © 2023
