Research on the Evaluation System of College Students' Innovation Ability Based on Neural Network

Innovation ability education is significant to individuals, teachers, and higher education institutions, and innovation ability assessment is key to understanding how instructional strategies affect the creative process and learner effectiveness. This study took computer major students as the research object, based on the multi-dimensional learning situation feature set of the university academic affairs office database. First, the innovation ability evaluation index system consisting of 15 indicators in three dimensions was constructed, and the entropy method was used to assign weights. Second, a reasonable evaluation of students' innovation ability was conducted based on a composite radial basis function–backpropagation neural network model, and the effectiveness of the model was also verified. Experiments show that, compared with a single radial basis function neural network and single backpropagation neural network, the average mean square error of the constructed model decreases by 22.83% and 12.35%, respectively. Its superior performance and strong generalization ability can be used to analyze problems related to the innovation ability of university and college students.