| Clustering analysis is the main research task of exploratory data analysis and an important direction in the field of data mining.The k-means algorithm is widely used in machine learning fields such as signal processing,image analysis,clustering model creation,and so on,and the robust correction of the k-means problem is also a very active research direction.It includes the k-means problem with penalties and thek-means problem with outliers.On the basis of studying robust deformation of k-means problem,combined with two types of robust robustness settings,the k-means problem is further extended,and the k-means problem with both punishment and outliers（-MPO）is considered.Finally,the algorithm with an approximate ratio of 274 is obtained.In chapter 1,firstly,the historical background and development process of thek-means clustering problem is introduced in detail.Secondly,the definitions,symbols,and necessary preparation knowledge of the k-means problem are introduced.In chapter 2,the running process of the k-means algorithm is introduced in detail at first,and an example is given.Then,we introduce the main problems of this paper.In chapter 3,firstly,the algorithm designed for k-means problem with penalties and outliers is presented.Then,the running process and the approximate ratio of the algorithm are as follows,and a simple implementation is given. |
