Data mining with K-means clustering using c

I expect you will find some useful information here[^].

Hi,

K- means clustering is a simple algorithm. Nothing much to worry. K stands for number of clusters. So here the number of clusters predefined.

So first take arbitrary means for each cluster expected. Then calculate the distance from that means to your data point and assign that point to the cluster which has minimal distance to the clusters mean. Once you have partitioned all the points recalculate the mean of each cluster and repeat the first step till the cluster mean not change from the accepted difference.

As this is iterative , instead of assign arbitrary means analyst the data histogram and take the modes of each group (if the data has clusters it will show multi modes) as the initial mean. In that way you can reduce the iterations and get faster response.

But yes, this algorithm goes complex if you have n-dimensional data. For a 2D data this very simple algorithm. Again yes, this algorithm improved a lot for a decade and now advanced algorithms available for improved accuracy. I given the basics of it, you can learn more and write your optimized codes now.