What is the minimum number of attribute in K-Means Clustering ?

<Contents | Previous | Next>

As you may guess, the minimum number of attribute is one. If the number of attribute is one, each example point represents a point in a distribution. The k-mean algorithm becomes the way to calculate the mean value of k distributions. Figure below is an example of distributions.

If you set k=1, you will get the mean (arithmetic average) of your data.

<Contents | Previous | Next>