Supervised Learning: Regression vs. Classification
Supervised learning algorithms learn from labeled data to predict either continuous numerical values (regression) or discrete categories (classification).
Core Principles
- Regression predicts a number, allowing for infinitely many possible outputs.
- Classification predicts categories, with a small, finite number of possible outputs.
Action Steps
- Identify whether your problem requires predicting a continuous value (regression) or a discrete category (classification).
- For classification, determine if the output is binary (e.g., benign/malignant) or multi-class (e.g., type 1/type 2/type 3).
Key Terms
- Supervised Learning: A type of machine learning where algorithms learn from a dataset containing 'right answers' (labeled data) to make predictions.
- Regression: A supervised learning task that predicts a continuous numerical value.
- Classification: A supervised learning task that predicts a discrete category or class label.
- Benign: In the context of medical diagnosis, a non-cancerous or non-dangerous condition.
- Malignant: In the context of medical diagnosis, a cancerous or dangerous condition.
Real World Examples
- Breast Cancer Detection: A classification problem where a machine learning model predicts if a tumor is 'benign' (0) or 'malignant' (1) based on features like tumor size and age.
- House Price Prediction: A regression problem where a model predicts the price of a house (a continuous number) based on features like size, location, and number of bedrooms.
- Image Recognition: A classification problem where a model predicts the category of an image, such as 'cat' or 'dog'.
Related cheat sheets
More like this