Research on the Prediction Model of Civic Education Achievement Based on the Combination of Artificial Intelligence Features

Under the situation of continuous expansion of opening up to the outside world, the ideological and political education work in China's colleges and universities is facing new situations and new requirements. In order to understand students' individual stage learning in time, improve the learning efficiency of ideological and political theory courses, and realize the prediction and early warning of students' course learning status, the study combines the disciplinary characteristics of ideological education and integrates the improved gradient boosting decision tree (GBDT) algorithm and the improved support vector machine (SVM) for feature combination of feature information for the construction of ideological education grade prediction model. The experiment demonstrates that the fineness of the improved particle parameter seeking model was increased, and the GBDT - SVM model performed optimally in terms of accuracy, precision, recall and F1, where the maximum accuracy was 92.13%, the highest precision was 93.70% and the maximum recall was 85.7%; the area under the curve values of the four algorithms were 0.95, 0.92, 0.81 and 0.73 respectively, and the algorithms have high generalization ability; on different data sets, the algorithms have the smallest average absolute percentage error and root mean square error values, and the comprehensive performance of the models turns out to be better. The research design model can accurately predict students' performance in civic education, which is conducive to teachers' teaching according to their ability.