Page 313 - Artificial Intellegence_v2.0_Class_11
P. 313

Hierarchical Clustering
                 Hierarchical clustering builds a tree of clusters. The aim of the algorithm is to produce a tiered series of nested
                 clusters. Each cluster is different from every other cluster, and the objects within each cluster are mostly similar to
                 each other.








                                                       E F                                                          F
                               A                 D                                                                  E
                                 B                                                                    D
                               C
                                                                                                                C
                                                                                                                 B

                                                                                                           A
                                                       I                                                          H

                                                     H
                                                            J                                                     I
                                           G                 K                                                    K
                                                                                                                  J
                                                                                                G






                        K-Means Clustering


                 Out of the various clustering techniques mentioned above, the simplest and very widely used clustering algorithm is
                 “centroid-based clustering using K-means”.
                 A centroid is an imaginary or real location denoting the centre of the cluster. The K-means algorithm identifies K
                 number of centroids, and then assigns every data point to the nearest cluster, while trying to keep the centroids as small
                 as possible. The algorithm has the following steps:

                 Step 1:   Decide the number of clusters (k)

                 Step 2:   Select k random points from the data as centroids

                 Step 3:   Group all the points to the nearest centroid
                 Step 4:   Calculate the centroid of newly formed clusters

                 Step 5:   Repeat steps 3 and 4

                 It is a repetitive process. It will keep on executing until there is no change in the centroids of newly formed clusters or
                 the maximum number of iterations are reached.









                                                                                    Classification & Clustering   311
   308   309   310   311   312   313   314   315   316   317   318