摘要：The purpose of educational data mining is to find the internal relations and laws hidden in the massive educational data and help students’ learning and teachers’ teaching and management. The prediction and analysis of student achievement can improve the way of training students and promote the improvement of teaching quality. This paper proposes a student grade prediction algorithm based on a DTGA-BP (decision tree genetic algorithm-back propagation) neural network to predict better and analyze students’ grades. Firstly, the algorithm preprocesses the data with a correlation analysis method to generate the initial population. Then, the feature selection of evaluation indexes is carried out through DTGA decision tree, and the number of hidden layer neurons is optimized. Finally, the crossover probability and mutation threshold of the BP neural network are used to optimize the initial weight. The experimental results show that the prediction results of this algorithm are more consistent with the actual results, more scientific and accurate than the traditional methods, and can provide better services for teaching and management.
