The Construction of Performance Prediction Model of Ideological and Political Education Based on Feature Extraction Algorithm

The this paper proposes a weighted ensemble learning algorithm with breadth learning algorithm as the individual learner. First, the breadth learning algorithm is used to predict students’ grades. Since the performance of a single classifier is not stable enough and is sensitive to data changes, the experiment uses the weighted integration method to improve the breadth learning algorithm; then, it is combined with the decision tree using weighted integration. The results of the algorithm and the multilayer perceptron neural network algorithm are compared. The research results can provide a method reference for student achievement prediction.