This page offers information on the Low Code Machine Learning course I initially taught at Shiraz University. The purpose of this course is to introduce the students to Machine Learning with the least programming and math prerequisites. We make use of Low Code ML tools such as PyCaret.
The approach of this course is the reverse of usual formal courses on ML. Instead of introducing the tools and then making use of them, I started with a general introduction to AI and ML and talked about technology democratization. We then solve five different types of problems: binary and multi-class classification, regression, clustering and time series forecasting. Each module of PyCaret, say the classification module, takes care of preprocessing and testing (e.g. cross validation) and compares all the available models based on relevant metrics. This way, starting from a real problem, the students are exposed to different models/methods and their evaluation metrics. I then explain to them the simpler methods and metrics.
Students are encouraged to run experiments of their own on the available datasets and to delve deeper into the math behind each model.
A guiding principle for this course is the following:
Learning ML should involve as few software and programming idiosyncrasies as possible.
The lecture notes for the course are available here. The jupyter notebooks will be available on Github.
See also this Linkedin post.