Articles → MACHINE LEARNING → Perceptron In Machine Learning

Perceptron In Machine Learning

This article describes a perceptron in Machine Learning.

Purpose

The perceptron takes multiple binary inputs and produces a single binary output.

The concept was to assign different weights to each input based on its importance, and then check whether the weighted sum exceeds a certain threshold. Only if it does, a decision such as yes is made; otherwise, the decision is no.

Example

Consider an example where you are planning to purchase a new smartphone.

Criteria	Input (xi = 0 or 1)	Weight (wi)
Phone is Affordable	x1 = 0 or 1	w1 = 0.8
Has a Better Camera	x2 = 0 or 1	w2 = 0.6
Battery Life is Good	x3 = 0 or 1	w3 = 0.7
Friends Recommend It	x4 = 0 or 1	w4 = 0.4
Attractive Design	x5 = 0 or 1	w5 = 0.3

Steps Of How A Perceptron Works

Set a threshold value
Multiply all inputs by their weights
Sum all the results
Activate the output

Set A Threshold Value

Threshold = 1.5.

Multiply All Inputs By Their Weights

Criteria	Input (xi = 0 or 1)	Weight (wi)	xi X wi
Phone is Affordable	x1 = 1	w1 = 0.8	0.8
Has a Better Camera	x2 = 0	w2 = 0.6	0
Battery Life is Good	x3 = 1	w3 = 0.7	0.7
Friends Recommend It	x4 = 0	w4 = 0.4	0
Attractive Design	x5 = 1	w5 = 0.3	0.3

Sum All The Results

0.8 + 0 + 0.7 + 0 + 0.3 = 1.8

Activate The Output

Return true if the sum is greater than 1.5 ("Yes, I will purchase the smartphone").

Transition And Rotation

Transition

Translation means shifting the line up, down, left, or right without changing its slope. In perceptron terms, this is controlled by the bias (b).

If you increase bias, the line moves upward.
If you decrease bias, the line moves downward.

Both lines are parallel (have the same slope) but are shifted.

Rotation

Rotation is changing the angle of the line by changing the weights.

Posted By -	Karan Gupta

Posted On -	Tuesday, September 30, 2025

Query/Feedback

Your Email Id		**

Subject		*

Query/Feedback	Characters remaining 250	**