Overview

Teaching: 5 min
Exercises: 0 min

Questions

• What are the major categories of machine learning?

• What are some example applications for each?

Objectives

• Understand the major categories of machine and what kind of problems they can each solve.

Major Categories of Machine Learning and Example Applications

Figure: Major categories of machine learning (image source: MathWorks).

1. Supervised Learning

Supervised learning learns a prediction model from past input data and observed outcomes, such that the prediction can take new input data and estimate outcome.

• Sub-categories of supervised learning:

  • Regression: outcomes are continuous values

    • Example applications:

      • Predict housing price given house size, location, number of rooms and year

      • Predict streamflow in the next week given mean annual streamflow, streamflow in the past week, rainfall in the past week and day of year

  • Classification: outcomes are categorical values

    • Example applications:

      • Classify handwritten digits

      • Land classification

Figure: Handwritten digits classification (image source: Wolfram).

Figure: Hurricane Harvey Houston flood mapping (upper panel: raw image; lower panel: flooding pixels classified).

2. Unsupervised Learning

Unsupervised learning groups, interprets and find patterns in input data without outcome data. For example, unsupervised learning can be applied to identify groups (or clusters) in the data.

Figure: Illustration of supervised classification vs. unsupervised clustering.

Key Points

• machine learning categories, supervised learning, unsupervised learning, regression, classification

---