A Novel Text Copy Detection Method based on Semantic Feature

Abstract

With the rapid development of the Internet, getting shared resource on the network is becoming more easy, and various plagiarize is becoming to breed, so the research of text copy detection technology is becoming more important. The traditional copy detection technology is based on term frequency statistics, and does not consider the context semantic. Some plagiarism can be easily made by replacing synonyms, changing the sentence structure, or translating from one language to another language. But the traditional copy detection technology could not detect such plagiarism. In this paper, a text copy detection method based on semantic is proposed. By using an improved TFIDF algorithm, terms could be more accurately extracted from each document in the corpus. By putting the documents corresponding to the terms one by one, a terms category is built in the database. When a document is detecting, the terms are read from the database and matched. The testing results show that, compared to the traditional TFIDF algorithm, the improved method could more accurately detect the plagiarism.

Ähnliche Arbeiten