Abstract
Monitoring and guiding instructional management require student performance evaluation. Traditional evaluation and analysis methods based on absolute scores, on the other hand, have certain flaws and are unable to fully reflect the information contained in student performance, thus limiting the impact of student performance evaluation on teaching and learning management. Data mining is regarded as the backbone technology for future information processing, and it introduces a new concept to the way humans use data. Schools must analyse and evaluate the performance of students in the same grade level and secondary school in a timely and staged manner. Clustering is a type of data mining that uses similarity rules to classify sample data into groups with a high degree of similarity. To address the difficulties caused by the wide variation in course difficulty in student performance evaluation, a method based on the K-means clustering algorithm is proposed. The K-means algorithm and the improved K-means algorithm with student information are investigated. The test results showed that the K-means clustering algorithm, the improved algorithm in this paper, and the fast global mean clustering algorithm all cluster the same randomly generated data set with noisy points, but the clustering time of the algorithm in this paper is only 0.04, which has obvious advantages. As a result, the clustering algorithm-based higher education management and student performance evaluation mechanism provides some insights for future research on student learning patterns. It is hoped that instructional administrators will gain a better understanding of students' learning characteristics so that they can better guide their teaching.