Novel Enthusiasm Evaluating Method for Code Programming Curriculum

In order to improve the logical ability of all youths, the governments of different countries have set the code programming curriculum in high school as one of their education policies. Hence, some programming aid software are developed for student trying to learn how to program, taking exams on the system, and evaluating the students according to the codes of students. However, it is hard to know the enthusiasm from learning of the students which directly impact the following learning status of students. Hence, observing the enthusiasm from learning of the students is important for teachers. In our previous work, we used the Fuzzy Logic to evaluate the enthusiasm of the students from the log files in the DICE system. However, it only observes the login frequency and during time. In this paper, we provide a novel way to improve the evaluating the enthusiasm of the students. First, we analyze the all the information from the log files. And then, we store all behaviors with time stamp of one student into one special tensor. After all tensors of all students ready, we send input them to the Convolution Neural Network (CNN) to classify the students to three categories “passion”, “normal” and “apathetic.” We used log files with the number of over 206. We separate the dataset into training and testing datasets 20 times randomly. The accuracy of training dataset is 100%, and the accuracy of average testing dataset is 93.01%. The experimental results show that we can separate the students into the three categories. Hence, for our new method can significantly measure the learning enthusiasm of the students.