Plagiarism in programming assignments has been increasing these days which affects the evaluation of students. Thispaper proposes a machine learning approach for plagiarism detection of programming assignments. Different features related to source code are computed based on similarity score of n-grams, code style similarity and dead codes. Then, xgboost model is used for training and predicting whether a pair of source code are plagiarised or not. Many plagiarism techniques ignores dead codes such as unused variables and functions in their predictions tasks. But number of unused variables and functions in the source code are considered in this paper. Using our features, the model achieved an accuracy score of 94% and average f1-score of 0.905 on the test set. We also compared the result of xgboost model with support vector machines(SVM) and report that xgboost model performed better on our dataset.
The Standard Coder: A Machine Learning Approach to Measuring the Effort Required to Produce Source Code Change
