AI + Python + Math Glossary

Every key term from this book, in one place.

Accuracy

Percentage of correct predictions. NOT always the right metric.

Algorithm

A step-by-step recipe for solving a problem.

Big-O

How an algorithm scales with input size — O(n), O(n²), etc.

Boolean

A True/False value. The atom of logic in any program.

Calibration

How honest a model’s confidence is. 90% sure → right 90% of the time.

Capstone

A final project showing everything you have learnt.

Cosine similarity

Compares two vectors by angle. Used everywhere in NLP.

CSV

Comma-Separated Values. The most common dataset file format.

DataFrame

A pandas table with rows and columns. The standard ML data structure.

Decision tree

A model that asks yes/no questions to reach a prediction.

Dot product

Vector × vector → number. Powers attention in LLMs.

Function

Reusable block of code. Built with def in Python.

Hackathon

Build-something-fast competition. Great way to ship your first AI.

Iris dataset

150 flowers with 4 measurements. The ‘Hello World’ of ML.

Joblib

Python library to save/load trained models as .pkl files.

Kaggle

Website hosting ML competitions and datasets — including Indian ones.

List

An ordered collection in Python: [1, 2, 3].

Mean

Sum divided by count. The most common ‘average’.

Median

Middle value when sorted. Resistant to outliers.

NumPy

Python’s numerical library. Arrays and fast maths operations.

Outlier

A value far from the others. Usually a typo or extreme case.

Pandas

Python library for tables. read_csv, head, describe, dropna.

Pickle (.pkl)

Python’s format for saving trained models to disk.

Probability

A number between 0 and 1 expressing how likely something is.

Random seed

A number that makes ‘random’ code reproducible across runs.

README

The text file that explains a project to anyone visiting your GitHub.

Scikit-learn

The standard Python ML library.

Standard deviation

How spread out the numbers are around the mean.

Train/test split

Divide data: ~80% for training, ~20% to test honestly.

Truth table

All combinations of True/False inputs for a Boolean operator.

Variable

A named storage for a value: age = 13. The first thing in Python.

Vector

A list of numbers. The basic unit of maths inside any AI.

Worked example

A problem with the full solution shown step by step.