Articles → MACHINE LEARNING → Terminologies Used In Machine Learning

Terminologies Used In Machine Learning

This article describes the different terminologies used in Machine Learning.

Mathematical Model

A mathematical model is a simplified representation of a real-world system using mathematical concepts, equations, and relationships. A common example of a mathematical model is the simple interest formula, which calculates the interest earned over time using a specific equation.

SI = (P × R × T) / 100

where P is the principal, R is the rate of interest, and T is time.

Labelled Dataset

If the dataset contains both inputs and the correct outputs (labels), it is called a labelled dataset. Consider the following example: -

Age	BMI	Blood Sugar	Label (Diabetes)
45	28.5	160	Yes
34	23	95	No
50	30.2	180	Yes
29	22.1	140	No

Each row is a patient record with input features (like age, BMI, and blood sugar) and a label indicating whether the person has diabetes.

Supervised Learning

Supervised learning is a machine learning type where the model is given input data and labels (output).

Unsupervised Learning

Unsupervised learning is a type of machine learning where the model is given input data without any labels (output). For example, you have the data on customers (age, spending habits, income), but no labels. You want to categorize them into natural groups (e.g., budget shoppers vs premium buyers).

Reinforcement Learning

Reinforcement Learning (RL) is a type of machine learning paradigm where an agent learns how to make decisions by interacting with an environment. Instead of being explicitly told what action to take (like in supervised learning), the agent learns through trial and error, guided by rewards or penalties.

Independent And Dependent Variables

In machine learning, independent variables (features) are the inputs used to predict the dependent variable (target/output).

Problem Types

Classification Type

In classification-type problems, we categorize the target into fixed categories. For example, classify the email as spam or not spam.

Regression Type

In regression-type problems, the target is continuous numerical values. It means that the output can take any value within a range (not limited to fixed categories or labels). There is no fixed or finite set of possible outputs.

If a regression problem involves predicting temperature and the expected values range from 10°C to 20°C, then the possible output range is 10 to 20. Any value within this range — such as 10.5, 11.6, or 19.99 — is valid. These values are continuous, meaning the model can predict any decimal or fractional number within that interval (not just whole numbers).

Posted By -	Karan Gupta

Posted On -	Friday, May 16, 2025

Updated On -	Monday, September 1, 2025

Query/Feedback

Your Email Id		**

Subject		*

Query/Feedback	Characters remaining 250	**