A Comparison of K-Means and Mean Shift Algorithms
Mehak Nigar Shumaila · International Journal of Theoretical and Applied Mathematics · 2021
Clustering, also known as cluster analysis, is a learning problem that occurs without the intervention of a human. This technique is frequently used very efficiently in data analysis to observe and identify interesting, useful, or desirable patterns in data. The clustering technique operates by dividing the data involved into similar objects based on their identified properties. This process results in the formation of groups, and each formed group is referred to as a cluster. A single said cluster consists of objects from the data that share similarities with other objects found in the same cluster and differ from objects identified from the data that now exist in other clusters. Clustering is an important process in many aspects of data analysis because it determines and presents the intrinsic grouping of objects in the data based on their attributes in a batch of unlabeled raw data. This method of cluster analysis lacks a textbook or, to put it another way, good criteria. This is due to the fact that this process is unique and customizable for each user who requires it for a variety of reasons. There is no single best clustering algorithm because it is so dependent on the user's scenario and needs. The purpose of this paper is to compare and contrast two different clustering algorithms. The algorithms under consideration are the k- mean and the mean shift. These algorithms are compared based on the following criteria: time complexity, training, prediction performance, and clustering algorithm accuracy.