‘K-means clustering’ is a type of data mining algorithm that literally involves the clustering of various observations into different groups. The best feature of k-means clustering is that the algorithm involved is very simple, and even without any knowledge of variable relationships, the group clustering may be done.  In the field of data mining, easy clustering provided by k-means results in a more efficient and faster way of determining patterns especially when the data or information involved is large.  With k-means clustering as the algorithm used, the extraction of data and analyzing them will be much easier, more efficient, and faster.

K-means clustering works through the creation of groups or so-called “k” clusters.  These clusters will then handle observation data and assign them into different groups based on how far or how near they are according to the mean data involved. With all these variables present, the process of clustering and computation will then commence. To simplify things, so-called “k” points are determined in the large set of data. Each k point will be measured according to its distance from the center.  The center will represent the mean value. When all the distances are now recorded, k-means clustering will now start, and this translates into the recomputation of the distance between the k points and the mean.  This algorithm cycle will go on until such time that the clustered groups are able to combine.  Converging of the clustered data or observations will then give way to faster and more efficient data analysis and/or management.

Various industries and segments of society have used k-means clustering as the algorithm of choice when it comes to their data mining or information management needs.  K-means clustering is considered very easy to use and implement and has been applied in various fields and sciences including agriculture, astronomy, and geostatistics, among many others.