Cluster Analysis of Student Scores Based on Global K-Means Algorithm

Aiming at the problem that the initial clustering center in the K-means algorithm is easily affected by outliers, it is proposed to analyze the data based on the global k-means algorithm. The global k-means algorithm is used to improve the determination process of the initial clustering center and reduce the influence of the random initial clustering center on the clustering result. First, preprocess the data of the four courses of the first semester of the 2018–2019 academic year for undergraduates, and save them in csv format and then, through the experimental comparison of the real traditional k-means algorithm and the global k-means algorithm, we get the clustering indicators such as Jaccard coefficient, accuracy, F value, etc. The experimental results show that the global k-means algorithm is 7.9% more accurate than the original k-means clustering algorithm, and it is verified that the global k-means clustering algorithm is better than the traditional k-means algorithm.