Introduction
This module explores a process called supervised learning, in which machine learning models learn from examples.
By understanding supervised learning, we start a deeper dive into the individual components of the learning process, and exactly how this process can improve a model. Through examples, we also explore how correctly setting up this learning process is critical to achieving a high-performance model.
Throughout this module, we use the following scenario to explain the process of supervised learning. This scenario provides an example for how you might meet these concepts while you're programming.
Your family has been managing Washington State's longest-running elk farm for several generations, but for decades the health of your herd has been slowly getting worse. It's a well known fact that your farm's breed of elk shouldn't be fed grain when average nightly temperatures are over freezing (32°F or 0°C). For that reason, you traditionally follow your grandfather's farming calendar and switch from grain feed after January 31.
You recently read about climate change affecting farming practices. Could this change explain the poorer health of your elk in recent years? With some historical weather data at your side, you seek to determine whether local temperatures are changed from your grandfather's day, and whether your farming calendar needs to be updated.
Prerequisites
You should have a basic familiarity with inputs, outputs, and models.
Learning objectives
In this module, you will:
- Define supervised and unsupervised learning.
- Explore how cost functions affect the learning process.
- Discover how to optimize models by using gradient descent.
- Experiment with learning rates, and see how they can affect training.