Exercise recommendation algorithm based on improved collaborative filtering
Zhizhuang Li, Haiyang Hu, Zhipeng Xia, Jianping Zhang, Xiaoli Li, Jingyan Shi, Hailong Li, Xuezhang Li · 2021
The recommendation method based on collaborative filtering has some shortcomings in the field of exercise recommendation, such as lack of interpretability and rationality. The existing methods for students' cognitive diagnosis are too rough to measure students' mastery of knowledge point, and the measurement of students' ability has the disadvantage of hysteresis. This paper proposes an exercise recommendation method aimed at improving students' mastery of the specified knowledge point faster. For the designated student and the designated knowledge point, this method can choose the exercise that can help student improve the level of mastery of the knowledge point as fast as possible in all the exercises including the knowledge point, and recommend to the student. This method is based on the improved cognitive diagnosis method and Long Short-term Memory Networks LSTM, and recommends exercises for the target students that can improve the knowledge level of students similar to the target students. According to the experimental test, the exercises recommended by this method can help target students to improve their mastery of the target knowledge point to a greater extent under the condition of doing the same number of exercises.