Introduction
Classification models’ outputs are categorical, meaning they can be used for labeling inputs or decision making. For example, a self-driving car uses classification to decide whether to turn left or right at a fork in the road. A classification model is different from classical regression models where outputs are continuous, such as the size of a shoe or the speed of a train. Classification models are diverse in how they work. To get started, let's focus on logistic regression, which is a simpler and popular type of model that is used extensively across many arms of science and industry.
Scenario: Predicting avalanches with machine learning
Throughout this module, we use the following example scenario to explain concepts related to classification. This scenario is designed to provide an example for how you might meet these concepts in your own programming.
Your charity is responsible for avalanche-rescue operations at hiking trails across the northwest of the United States. Granted, the safest option would be to permanently close all trails during skiing and hiking season but that would mean no sportspeople would get to enjoy the great outdoors! Your goal is to build a model that can predict whether an individual day is likely to result in an avalanche. Then by using that prediction, you can close the trail when the risk is high. Keep in mind as you make predictions: Predicting avalanches that don't happen can hurt local tourism while failing to predict avalanches that do happen can result in loss of life. Clearly, a balance must be found.
Caution
The data for these exercises are fabricated and are solely for educational purposes. For those eager hikers and skiers out there: Machine learning can be used for avalanche prediction but don’t use this data or your trained model for anything except learning about machine learning.
Prerequisites
- Familiarity with machine learning models
Learning objectives
In this module, you will:
- Discover how classification differs from classical regression
- Build models that can perform classification tasks
- Explore how to assess and improve classification models