Classification

It is a supervised learning approach. In classification, data points are classified based on some characteristics features. Different classification algorithms are Decision tree, Naïve Bayes, Linear Discriminant Analysis, K-Nearest Neighbour, Support Vector Machines(SVM).

K-NN: It is the most common method of classification. We simply calculate the Euclidian distance between the data points, find the nearest points, and classify the data point to the nearest class. K determines the number of data points nearest to the given point.

Let’s take an example, we are given data points of class 1 and class 2. We are asked to find the

Class of the points A (6,11) and B  (14,3)

Class 1	Class 2	Distance of point  (6,11) from Class 1                 class 2		Distance of point  (14,3) from Class 1             class 2
(11,11)	(7,11)	5	8.54	1	10.36
(13,11)	(15,9)	7	8.06	9.21	6.08
(8,10)	(15,7)	2.23	9.21	9.84	4.12
(9,9)	(13,5)	3.60	7.81	9.21	2.23
(7,7)	(14,4)	4.12	8.06	10.63	1
(7,5)	(9,3)	6.08	7.28	8.54	5
(15,3)	(11,3)	12.04	1	9.43	3

If we assume K = 1, one nearest data point, then the A can be classified in class 2 as it’s shortest distance is 1 while B will be classified to both the classes.

For K=2, two nearest data points, A will be classified into class1, and B will be classified into class 2.