Provide a new method of clustering using k -median strategies for the replace of K -Means methods

One of the most important activities in data analysis is dividing data into a set of categories or clusters. Clustering is the grouping of a particular set of elements based on their similarities, there are several different ways to implement this partitioning, such as distinct models. distinct models are distinguished by their organization and the type of relationship between them. In this article, we present the Mean algorithm based on clustering and compare it with the k -means algorithm, on the other hand, after evaluating the clusters also the median of the implementation of the clusters has been evaluated and compared for the best result. Mean algorithm shows that clustering by the median of the data in groups is more appropriate than obtaining their average. Furthermore, at the end of the implementation we used DBSCAN algorithm, the DBSCAN algorithm allows us to find an element that is not part of any cluster, or that belongs entirely to it.

Keywords:

Clustering , Data mining , Median , k -means , DBSCAN Algorithm