Achievement prediction and analysis based on neural network for smart education

In the traditional teaching mode, it is difficult for teachers to have a comprehensive understanding of each student’s study, and it is also hard for them to provide targeted guidance and assistance. With the development of data collection and analysis technology, schools and educational institutions can make better use of big data technology to analyze students' learning data and predict their academic performance. In order that teachers can better understand the learning characteristics and needs of each student to realize personalized teaching, this paper investigates the impact of five usual performances, including students' attendance, homework, topic report, communication, and answering questions, on students' final exam results. In this paper, we select the usual scores and final scores of 225 students in a university, and use BP neural network to analyze the relationship between these data, establish a prediction model, and compare and analyze the actual scores of students at the end of the term with the predicted results through various aspects. Then, the K-Fold cross-validation method was used to compare the students' actual scores and predicted scores at the end of the semester. The results show that the BP neural network model can effectively predict students' final results and promote targeted personalized education.