Fundamentals of Supervised Machine Learning

 

Fundamentals of Supervised Machine Learning

Supervised machine learning is a type of machine learning where algorithms learn from labeled data—datasets in which each input example is paired with the correct output (label)—so the model can learn the mapping from inputs to outputs and make accurate predictions on new, unseen data.

Core idea

• Labeled training data: Each training example has features (inputs, often denoted X) and a target/label (output, y)
• Learning objective: Find a function f such that y \approx f(X), minimizing prediction error on new data.
• “Supervised”: The labels act like a teacher, telling the algorithm what the correct answer should be for each input.

Main problem types

Supervised learning is typically divided into two broad task types:

• Classification
• Output is a category or class label.
• Examples: spam vs. not spam email; disease present vs. absent; image of a digit 0–9
• Regression
• Output is a continuous numeric value.
• Examples: house price prediction; temperature forecasting; stock price estimation.

Common algorithms

Typical supervised learning algorithms include;

Linear and logistic regression

• Decision trees and tree ensembles (random forests, gradient boosting)
• Support vector machines (SVM)
• k-nearest neighbors (k-NN)
• Neural networks (including deep learning models)

These differ in how they model the relationship between X and y, their assumptions, and their suitability for different data sizes and types

Typical workflow

A standard supervised learning pipeline looks like this:

1. Collect and label data
   Gather examples where both inputs and correct outputs are known.
2. Split data
   Divide into training, validation, and test sets (e.g., 80/10/10).
3. Choose a model and train
   Fit the algorithm on the training data to learn f.
4. Evaluate
   Measure performance on validation/test data using metrics like accuracy, F1 score (classification) or RMSE, MAE (regression).
5. Deploy and monitor
   Use the model for inference on new data and track performance over time, retraining as needed.

How it differs from unsupervised learning

• Supervised: Uses labeled data to predict specific outcomes (classification/regression)
• Unsupervised: Uses unlabeled data to discover structure (e.g., clustering, dimensionality reduction) without predefined targets.