Machine Learning

This lesson introduces the fundamental concepts of machine learning using different types of candies.

It is based on the lesson template used in Data Carpentry and Software Carpentry workshops.

Prerequisites

Schedule

	Setup	Download files required for the lesson
00:00	1. Introduction	What is machine learning? Where is machine learning useful?
00:25	2. Problem Set-up: Classifying Candy	We have a box of candies all mixed up together. What now
00:30	3. Incorporating Ethics into your Machine Learning Project	Is the project I am planning to do an ethical application of machine learning? Should I do this project in the first place? If I do this project, who might be harmed by it? How could the methods/tools/software I’m developing be misused for purposes I consider unethical?
01:10	4. Feature Engineering	Can a computer classify candy as easily as a computer does? How can we derive meaningful summaries of information from our candy to use in classification? Which features are better suited for classification, which are not?
01:40	5. Decision Boundaries	How can we tell whether the features we have recorded for our samples are good at separating the different classes? How can we build visual representations of our features?
02:00	6. K-Nearest Neighbours	How can we get a computer to classify new objects using the training data we recorded?
02:30	7. Model Evaluation	How can we evaluate whether our model does a good job? How can we define what a ‘good job’ means to us when setting up a machine learning model to solve a problem?
03:00	8. Logistic Regression	How can we get a computer to draw decision boundaries between different types of candies?
03:40	9. Cross-Validation	How can we determine whether a model generalizes to new data? How do we diagnose over-/underfitting in our model?
04:00	10. Final Thoughts
04:15	Finish

The actual schedule may vary slightly depending on the topics and exercises chosen by the instructor.