About this course
Machine learning is becoming more and more popular as a tool for analysis. But there is a big gap between knowing a little bit about machine learning algorithms to being able to get the most out of these techniques in a business.
This is a guided practical course to give you experience in applying machine learning algorithms for applications like forecasting.
Hierarchical (or multi-level) models provide a way of including and clarifying assumptions when trying to understand data. For example, you might want to assume that all products in a particular category are similar or that all variations in a split test will have something in common.
Combined with a Bayesian approach this gives a flexible way of building and validating models with many applications for digital businesses. One such application is in forecasting where the ability to include business specific assumptions in the model can make a big difference. Usual methods of calculating seasonality struggle with movable feasts (like Easter) let alone more complicated issues such as new product launches or stock shortages.
This is an advanced course, and attendees should have some familiarity with either R or Python. You will require a laptop with R, R Studio, Stan and Shinystan to fully participate in the course.