Question

Different clustering algorithms behave different on the same dataset. Single Link
and Complete Link are two agglomerative clustering algorithms.
Find a data file D such that these two algorithms behave differently on D, and
display their behavior. (A set of 6 to 10 points probably works well.)

Answer 1

In both clustering algorithm, in each iteration we merge two clusters having minimum distance. The difference is how we compute distance between two clusters.

In single linkage, distance between two clusters is the minimum of the pairwise distances between each points of two clusters. i.e. distance between closest elements in clusters.

In complete linkage, distance between two clusters is the distance between farthest elements of clusters. i.e. maximum of pairwise distances.

The following example demonstrates the difference:

be: 1 2356 0 6 59 0 5 11 lo 2 & 6 9 8 10 8 0 Coue1: mih de's tountei4 2.st (3,5)2. So, cluster "35" is tie加sL clus ttr 35 1 24 6 35 0 dist (35,C) min (dist ( 3,c) 4 8 6 5 61 8 9 8 80 dust Us, c)) 135 2 4 125 0 2. 4- 6 0 8

Date Page 1352 4 6 13 5 0 35 24m 12345 1234-5 1 23 4-S 0 d n

Date Page 3 and 5 dist (35, c)max (dsst (,c), dist (,) 35 1 24G 1 11 O 4-19650 6108 &o 35 | O 2. 2 10O Naw ctuut 35 2.4- ו6 35 O 24 t0 囗 6 t0 89 om "24-6 35 246 1 35e 24610 90

Date: Pege 1 to tom 1 244 1246 3S 124-1 3 S 6 1 1 6 2 3 5
please provide positive feedback if it helped.