Learn Machine Learning

What Is Machine Learning?

Machine Learning (ML) is a branch of artificial intelligence that lets a computer program find patterns in data and make predictions or decisions without being explicitly programmed for every scenario. In Python, ML relies heavily on NumPy for fast array math and the built-in statistics module for quick summaries, so almost every ML pipeline starts by loading data into arrays and computing basic statistics before any model is trained.

Traditional Programming vs Machine Learning

In traditional programming, a developer writes explicit rules that transform input into output. In machine learning the roles reverse: we provide input data together with known outputs, and the algorithm works backward to discover the rule that connects them. That discovered rule, called a model, can then predict outputs for new, unseen inputs.

Supervised Learning

Supervised learning trains a model on labeled data, meaning every training example comes with the correct answer attached. If the label is a number the task is called regression (predicting house prices, for example); if the label is a category the task is called classification (spam vs. not spam). The model learns the mapping from features to labels by minimizing prediction error on the training examples.

import numpy as np
import statistics

# Features: hours of daily study, Label: exam score (a supervised learning setup)
study_hours = np.array([1, 2, 3, 4, 5, 6, 7, 8])
exam_scores = np.array([52, 58, 63, 69, 74, 81, 88, 93])

print('Average study hours:', statistics.mean(study_hours))
print('Average exam score:', statistics.mean(exam_scores))
print('Correlation between hours and scores:', np.corrcoef(study_hours, exam_scores)[0, 1])

Unsupervised Learning

Unsupervised learning works with unlabeled data, so there is no correct answer to check against. Instead the algorithm looks for hidden structure on its own, such as grouping similar data points together (clustering) or reducing the number of variables while keeping the important patterns (dimensionality reduction). It is often used to explore data before building a supervised model.

import numpy as np
import statistics

# No labels here - just raw purchase amounts to explore for structure
purchase_amounts = np.array([12, 15, 250, 14, 260, 13, 245, 16, 255, 11])

midpoint = statistics.median(purchase_amounts)
small_purchases = purchase_amounts[purchase_amounts < midpoint]
large_purchases = purchase_amounts[purchase_amounts >= midpoint]

print('Median purchase amount:', midpoint)
print('Small-purchase group:', small_purchases)
print('Large-purchase group:', large_purchases)

• Data collection - gather raw numbers, text, or images relevant to the problem
• Data preparation - clean, organize, and convert data into numeric arrays
• Model training - let an algorithm find the pattern in the data
• Evaluation - measure how well the model performs on new, held-out data
• Prediction - use the trained model on new, unseen inputs

import numpy as np

np.random.seed(1)

# Simulate a tiny labeled dataset: house size (sqm) -> price ($1000s)
house_size = np.array([50, 60, 70, 80, 90, 100])
price = np.array([150, 175, 200, 225, 250, 275])

# A model is just a rule connecting inputs to outputs - here, a simple ratio
price_per_sqm = np.mean(price / house_size)
predicted_price_for_75sqm = price_per_sqm * 75

print('Learned price per sqm:', price_per_sqm)
print('Predicted price for a 75 sqm house:', predicted_price_for_75sqm)