Test Cheating Detection Method Based on Random Forest

Students cheating in exams destroys the fair principle of evaluation and affects the normal teaching order of the school. Therefore, the examination cheating detection has the vital significance. The existing cheating detection methods have disadvantages such as insufficient modeling accuracy for students, lag in cognitive diagnosis, difficulty in detecting multi-source plagiarism and low accuracy. In order to solve the shortcomings of the existing test cheating detection methods, this paper proposes a cheating detection method based on random forest. For a specific exam question, we find out which exercises the student has done in the usual practice with the same knowledge point as the exam question. We use the student's right or wrong of these exercises as the eigenvalues, establish a random forest model, and predict whether the students can correctly answer the questions in the exam. After the establishment of the random forest model, we used the random forest model to predict the student's scores for each question and compare them to the student's actual score on the exam. Finally, we judge whether the students cheat or not according to the gap between the predicted score and the real score and the similarity between the students and the test papers of surrounding students. Through experimental tests, the accuracy rate and recall rate of this method are significantly higher than the commonly used cheating detection methods based on test paper similarity and personal fitting index.