Page 344 - AI Ver 3.0 Class 11
P. 344
• Imbalanced classification: Imbalanced classification involves tasks
associated with classification in which the distribution of classes in the
dataset is strongly distorted, implying that a single class (the majority
x
2
class) surpasses the other(s). This could lead to problems throughout
training since the model may get biased against the majority class while
doing negatively on the minority class.
x 1
K-Nearest Neighbor(KNN) Algorithm
K-Nearest Neighbor (KNN) is one of the most basic yet essential classification algorithms in machine learning. It is a
simple, straightforward, and adaptable machine learning method. The KNN algorithm is a classifier using supervised
learning and non-parametric learning (makes no assumptions about the original data distribution) that employs
proximity or closeness to classify or forecast the arrangement of a single data point.
KNN is most beneficial where labelled data is prohibitively expensive or difficult to gather and it can perform well in
a wide range of forecasting situations. It has a wide range of applications, including handwriting detection, picture
recognition, pattern recognition and video recognition.
While the KNN technique can be applied to both regression and classification problems, it is most commonly employed
for classification, with the assumption that similar data points are possible to find close proximity.
Why is KNN Algorithm Required?
Assume you have a collection of coloured balls divided into three colours: red, blue, and green. If you purchase a new
ball but do not specify its colour, you will be unable to determine which colour group it belongs to. In such cases,
when the decision limits are unclear or the dataset lacks a well-defined structure, the KNN algorithm is applied.
The K-Nearest Neighbor (KNN) method determines the colour of the ball based on proximity of new ball or datapoint.
If the new ball is close to the red colour group, it is labelled as red; if it close to the blue colour group, it is labelled as
blue; and if it close to the green colour group, it is labelled as green.
x 2 Before K-NN x 2 After K-NN
Category B Category B
Category C Category C
K-NN
New data point New data point
assigned to
Category B
Category A Category A
x 1 x 1
342 Touchpad Artificial Intelligence (Ver. 3.0)-XI

