A Study on the Application of optimized K-Means Algorithm in Intelligent Hospital Procurement Supplier Management

In the procurement supplier management system of intelligent hospitals, the K-means clustering algorithm is widely used for clustering analysis. However, the K-means algorithm is sensitive to the selection of initial clustering centers and may generate local optimal solutions or fall into local minima. So this research improved the traditional K-mean clustering algorithm by combining the effectiveness evaluation index of K-mean clustering algorithm with particle swarm algorithm. And on the basis of the improved K-mean clustering algorithm, this research designs the supplier selection method of raw material intelligent procurement system. The experimental results show that the improved K-mean clustering algorithm has the highest initial profile coefficient of 0.51, and its profile coefficient is the highest when the number of samples reaches 300, which is 0.79. The improved K-mean clustering algorithm’s initial Dunn’s index and its final Dunn’s index are higher than the other three comparative algorithms, which are 59 and 67, respectively, and its Calinski-Harabasz index is the highest on average at 37.6, ranging from 3.4 to 11.2 higher than the other three compared algorithms. The optimized K-mean clustering algorithm has a larger Calinski-Harabasz index, which indicates that the inter-cluster variance is larger, the intra-cluster variance is smaller, and the clustering results are relatively better. This study is effective for the design of intelligent hospital purchasing supplier management system based on optimized K-mean clustering algorithm, which can be used as a technical reference for the customer segmentation problem of clustering algorithm.