K-means algorithm

The K-means algorithm is simple, it belongs to the unsupervised learning algorithm in the clustering algorithm, the use of European distance to the aggregation.

Solve the problem ha: there are a bunch of non-tagged training samples, and they can potentially be divided into K-class, how do we divide them? Then we'll use the K-means algorithm to divide it.

The algorithm is simple, so you can do it:

First step: Randomly initialize the center point of each category, U1,u2,u3,......, UK;

Step two: Repeat the following procedure:

Then, there is no then, and so on.

Too simple to explain.

K-means algorithm