class4 - now for K-means 1 set the number of clusrers =k 2)...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
Sheet1 Page 1 how to find the Z score TR190 13000 TR2500 4000 TR3700 25000 Mean430 14000 Std dev 310 10535.65 fyi small sample set === (<30) little sigma std dev for whole population S stddev for sample set zscore = (measured-mean) over stddev 1) MST start off with each node is its own cluster 2) find the centroid of clusters 3) merge the two closest clusters 4) decress #cluster by 1 5) goto 2 unless #clusters is small enough
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: now for K-means 1 set the number of clusrers =k 2) choose k starting points ato be used as initial centroids (1st k point, last k, point k, random points) 3) examine each point and allocate it to the clusrter whose centroid is nearest. 4) recalculate the cluster centroid after each addition Homework 158 34 128 *015 23 33 *017 26 *019 *019 *016 do an MST and a Kmeans (with 4) due tuesday...
View Full Document

Ask a homework question - tutors are online