Owing to K-means algorithm has the shortcoming that it always neglects the influence of cluster size when the Euclidean distances between samples and cluster center is calculated. In order to overcome the lack, the influence of cluster size is introduced into K-means algorithm in this paper. Therefore an improved K-means algorithm based on gravity is proposed, namely GK-means algorithm. The experimental simulation results show that GK-means algorithm has better performance compared with K-means algorithm. So the GK-means algorithm is adopted for assessing the performance of culture industry listed companies in this paper. Furthermore some satisfactory results are also obtained.
Sun, Haibo; Wang, Limin; and Han, Xuming, "An Improved K-means Algorithm and Its Application for Assessment of Culture Industry Listed Companies" (2015). WHICEB 2015 Proceedings. 4.