Alberta Machine Intelligence Institute - Machine Learning Algorithms: Supervised Learning Tip to Tail
This course takes you from understanding the fundamentals of a machine learning project. Learners will understand and implement supervised learning techniques on real case studies to analyze business case scenarios where decision trees, k-nearest neighbours and support vector machines are optimally used. Learners will also gain skills to contrast the practical consequences of different data preparation steps and describe common production issues in applied ML.
This is the second course of the Applied Machine Learning Specialization brought to you by Coursera and the Alberta Machine Intelligence Institute.
- Welcome to Supervised Learning, Tip to Tail! This week we'll go over the basics of supervised learning, particularly classification, as well as teach you about two classification algorithms: decision trees and k-NN. You'll get started programming on the platform through Jupyter notebooks and start to familiarize yourself with all the issues that arise when using machine learning for classification.
- Welcome to the second week of the course! In this week you'll learn all about regression algorithms, the other side of supervised learning. We'll introduce you to the idea of finding lines, optimization criteria, and all the associated issues. Through regression we'll see the interactions between model complexity and accuracy, and you'll get a first taste of how regression and classification might relate.
- This week we'll be diving straight in to using regression for classification. We'll describe all the fundamental pieces that make up the support vector machine algorithms, so that you can understand how many seemingly unrelated machine learning algorithms tie together. We'll introduce you to logistic regression, neural networks, and support vector machines, and show you how to implement two of those.
- Now at the tail end of the course, we're going to go over how to know how well your model is actually performing and what you can do to get even better performance from it. We'll review assessment questions particular to regression and classification, and introduce some other tools that really help you analyze your model performance. The topics covered this week aim to give you confidence in your models, so you're ready to unlock the power of machine learning for your business goals.