k-means clustering calculator

Do you want to find out what structure is present in your data? Then use the Cluster Analysis calculator here at DATAtab and select at least two metric variables. With cluster analyses, you can discover similarity structures in your data.

Do you want to calculate a cluster analysis? Only three steps are necessary:

Copy your data into the table
Select more than one variable
Select the number of clusters you want to calculate

Clusters can be calculated using various grouping methods. These can be divided into

graph-theoretical
hierarchically
partitioning
optimizing

DATAtab calculates you the k-means Cluster and hierachical cluster. The algorithm assigns each data point to the cluster whose center (or "centroid") is closest to it. The centroids are recalculated after each assignment, and the process is repeated until the clusters no longer change significantly. This helps to identify patterns or structure in the data.

k means calculator online

The k-Means method, which was developed by MacQueen (1967), is one of the most widely used non-hierarchical methods. It is a partitioning method, which is particularly suitable for large amounts of data.

First, an initial partition with k clusters (given number of clusters) is created.
Then, starting with the first object in the first cluster, Euclidean distances of all objects to all cluster foci are calculated.
If an object is detected whose distance to the center of gravity of the own cluster is greater than the distance to the center of gravity (centroid) of another cluster, this object is shifted to the other cluster.
Finally, the centroids of the two changed clusters are calculated again, since the compositions have changed here.
These steps are repeated until each object is located in a cluster with the smallest distance to its centroid (center of the cluster) (optimal solution).

Optimal cluster number

The number of clusters in the k-Means method must be determined before the start and is therefore not determined by the cluster method. But what is the optimal number of clusters in the k-Means method? The elbow method is a common way to determine the appropriate number of clusters.

Elbow curve

When you want to calculate a cluster analysis, often the big question is how many clusters should I take, The Elbow Method helps with this question! With each new cluster, the total variation in each cluster becomes smaller and smaller. In the extreme case, when there are as many clusters as there are points, the result is zero. However, in most cases, the reduction of the total variation becomes smaller after a certain point. This point is then used as the optimal cluster number.

Scaling data for k-means clustering

If the variables under consideration do not have the same unit, it is often advisable to scale the data before cluster analysis.