Detection of Plagiarism in Student Essays

This paper presents two methods for automatic detection of plagiarism in student essays, using Dutch text corpora to show their effectiveness. The first method is based on measuring the overlap in word trigrams between two essays, excluding all trigrams from the assignment text. This method proves efficient and robust, but relies on the availability of the plagiarized source. The second method compares an essay’s linguistic profile to the profile of the claimed author’s previous written texts. This approach proves promising, but as yet not completely able to detect all cases of plagiarism.

• Type: Conference
• Year: 2003
• Where: CLIN
• Authors: Hans van Halteren
• Keywords: CLIN 2003, Computational Linguistics, Dutch Text Corpora, Word Trigrams, first Method