增量k-Median算法的研究与实现

he paper focuses on the application of dynamic data clustering analysis, and then it makes a deep research. The paper first discusses the uncapacitated incremental facility location problem, and based on the concepts of neighborhood and potential it gives a reasonable solution and implements the IFL algorithm. Based on the IFL algorithm the paper solves the incremental k-Median problem through merging clusters halfway, and implements the IM(k) algorithm. At last, through simulating the IM(k) algorithm and doing a comparison test among the algorithm and other clustering algorithms, we prove that the IM(k) algorithms can deal with large datasets in the incremental problem, find its shortages and present its improvement direction.