Page 352 - AI_Ver_3.0_class_11
P. 352
K-Means Clustering
Randomly Select Each Object Assigned Clusters Centres Updated
K-Clusters (K=2) To Similar Centroid Depending On Renewed Cluster
Mean
100 100 100
90 90 90
80 80 80
70 70 70
60 60 60
50 50 50
40 40 40
30 30 30
20 20 20
10 10 10
0 0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Re-Assign Re-Assign
Data Points Data Points
Update Cluster
100 Centres 100
90 90
80 80
70 70
60 60
50 50
40 40
30 30
20 20
10 Iterative 10
0 0
0 10 20 30 40 50 60 70 80 90 100 Process 0 10 20 30 40 50 60 70 80 90 100
Advantages of K-Means Clustering
Some of the advantages of K-Means Clustering are:
• Easy to implement.
• Can handle large data sets.
• Can give initial positions to centroids (randomly).
• Easily adapts to new data.
• Can easily adapt to clusters of different shapes and sizes, like elliptical clusters.
Disadvantages of K-Means Clustering
Some of the disadvantages of K-Means Clustering are:
• K has to be chosen manually and it is not an easy process.
• The algorithm is dependent on initial values.
• Outliers greatly affect the clustering process.
• The algorithm has trouble grouping data where clusters are of fluctuating sizes and density.
Why is Clustering Unsupervised?
Clustering is an unsupervised machine learning technique that automatically divides the data into clusters or groups
of similar elements. The algorithm does this without any knowledge of how the groups should look in advance. So,
clustering is rather used for the discovery of knowledge rather than for prediction. It provides an idea of natural
groupings that are within data.
350 Touchpad Artificial Intelligence (Ver. 3.0)-XI

