High obfuscation plagiarism detection using multi-feature fusion based on Logical Regression model

The identification of high-obfuscation plagiarism seeds is one of the most difficult problems to be solved in plagiarism detection. Single feature type cannot identify the plagiarism seeds effectively because of the varied plagiarism methods used in high-obfuscation plagiarism. In this paper, a multi-features fusion method based on Logical Regression model for the high-obfuscation plagiarism seeds identification was proposed. This method used Logical Regression model to combine lexicon features, syntax features, semantics features and structure features extracted from suspicious text fragments pairs. Experiments show that the method is feasible and effective.