Exploring Machine Learning in Action

Regression

Regression is a supervised learning technique used when the target variable is a real number. The model learns the mathematical relationship between independent variables (features) and a dependent variable (target) to predict new outcomes.

How it works:

• Input: Features such as apartment size, postal code, number of bedrooms.
• Output: A numeric prediction (e.g., $1,200 monthly rent).
• The model fits a function (often a line, curve, or more complex equation) to best approximate the relationship in the training data.

Real-world applications:

• Predicting house prices
• Forecasting sales and revenue
• Estimating delivery or waiting times

Try it here: aka.ms/rent-predictor

Classification

Definition: Classification is a supervised learning technique where the target variable is categorical (discrete labels or classes). The model learns decision boundaries that separate data into these predefined categories.

How it works:

• Input: Features such as seed length, width, groove length.
• Output: A class label (e.g., Kama, Rosa, or Canadian wheat).
• The model analyzes training data and learns the rules/patterns that map inputs to categories.

Real-world applications:

• Email filtering: spam vs. not spam
• Medical diagnosis: benign vs. malignant tumor
• Sentiment analysis: positive, neutral, negative

Try it here: aka.ms/seed-identifier

Clustering

Definition: Clustering is an unsupervised learning technique, meaning the target variable is unknown. Instead of predicting labels, the algorithm groups data into clusters based on similarity patterns.

How it works:

• Input: Features such as purchase frequency and average spend.
• Output: Clusters (groups) of customers with similar behavior.
• The algorithm measures distances (e.g., Euclidean distance) and groups closer points together.

Real-world applications:

• Customer segmentation for marketing
• Document/news article grouping
• Image compression or anomaly detection

Try it here: aka.ms/customer-segmentation