Feature Engineering Notes
Subject Overview
The first course in the AI/
Subject at a glance
2
Units
12
Topics
2
Credits
2
Infographics
Unit-wise Syllabus
2 units — open any unit for detailed topic-wise notes, diagrams and important exam questions
- Importance of features in machine learning
- data and feature types (numerical, categorical, ordinal, discrete, continuous, interval, ratio)
- basic preprocessing — handling missing data
- data cleaning
- feature scaling
- normalization and transformation
Numerical: discrete
Counts, e.g. number of orders
Numerical: continuous
Measurements, e.g. height
Categorical: nominal
No order, e.g. city
Categorical: ordinal
Ordered, e.g. low/medium/high
Study guide
A machine learning model is only as good as the features it learns from. You learn why features matter, the types of data (numerical, categorical, ordinal, discrete, continuous, interval, ratio) and basic preprocessing: handling missing data, cleaning, scaling, normalisation and transformation. Practise identifying the type of each column in a data set; it decides which technique to use.
- Binning and discretization
- polynomial and interaction features
- one-hot and label encoding
- feature extraction vs feature selection
- filter/wrapper/hybrid selection methods
- feature reduction via Principal Component Analysis
How
A new 0/1 column per category
Each category gets an integer
Implies order
No
Yes
Best for
Nominal data such as colour
Ordinal data such as size
Drawback
Many columns for many categories
Models may read false order into nominal data
Study guide
This unit covers the techniques that improve models: binning and discretisation, polynomial and interaction features, one-hot and label encoding, feature extraction versus feature selection, filter, wrapper and hybrid selection methods, and dimensionality reduction with Principal Component Analysis (PCA). Know when to use one-hot encoding (no order) versus label encoding (ordered categories).
Why Feature Engineering Matters
Experienced data scientists often say that better features beat fancier algorithms. This paper teaches the preparation work that decides whether a machine learning project succeeds, and it is the bridge from basic analytics to AI.
How to Prepare for Feature Engineering
- Practise on one real data set: clean it, scale it, encode it and select features.
- Learn the difference between feature extraction and feature selection.
- Understand what PCA does in simple words, reducing many variables to a few that keep most of the information.
Feature Engineering: Frequently Asked Questions
What is feature engineering?
It is the process of creating, transforming and selecting input variables from raw data so that a machine learning model can learn from it more effectively.
What is the difference between normalisation and scaling?
Both bring numerical features to a comparable range; normalisation typically rescales values to a fixed range such as 0 to 1, while standardisation scales to zero mean and unit variance.