CLUSTER.KMEANS.CENTROIDS
Description
Clustering is a data mining technique for grouping a set of objects into smaller
groups, where each group's members are similar to the other members in some aspect.
Clustering is used in machine learning, for example, where the goal is to find
meaningful structures or to explain processes. You can cluster quantitatively using
numbers, or qualitatively using categories. CLUSTER.KMEANS.CENTROIDS is a
quantitative function that enables you to cluster a range of data in a workbook, by
locating the center point of each group and computing the distances between points
and group centers. The function groups each data point with the cluster having the
nearest center.
Syntax
CLUSTER.KMEANS.CENTROIDS(
range1
,
k
, [
maxIterations
], [
distanceMeasure
], [
emptyStrategy
])
- range1: The range of cells to analyze.
- k: The number of clusters to make.
- maxIterations: The maximum number of times to re-run the algorithm. If you don't specify a value, the limit is 100.
- distanceMeasure: The algorithm used to find the distance between points. Possible values are CanberraDistance, ChebyshevDistance, EarthMoversDistance, EuclideanDistance, or ManhattanDistance. If you don't specify a value, the function uses EuclideanDistance.
- emptyStrategy: The strategy to use if the function finds empty clusters while running the algorithm iterations. Possible values are ERROR, FARTHEST_POINT, LARGEST_POINTS_NUMBER, or LARGEST_VARIANCE. If you don't specify a value, the function uses LARGEST_VARIANCE. If you specify ERROR, the function returns #ERROR in the cell with a description. Example: CLUSTER.KMEANS: Empty cluster in k-means.