ML Resources

There is an incredible amount of information online. I will only refer to material that I have personally read or watched. See posts on Hacker News or Stack Overflow for others’ perspectives.

Download the texts you use to your computer and flash drive!

Texts: Primary

  • Introduction to Statistical Learning, a.k.a ISLR: The first text for the course. Lots of practice applying techniques on practical datasets. The book shows how to do the analysis in the R language, but we will be working in Python and implementing the algorithms on our own where it is practical.

Videos

  • Andrew Ng of Stanford wrote a course in Machine Learning: Overview without lots of mathematical details. Great for watching individually or on days the teacher is absent. Environment: free Octave or commercial Matlab. If you want to experiment, use Octave Online. Although this course seems light on the mathematics, reviewers with serious mathematical skills say that it gave them the overview to understand machine learning.

  • Yaser Abu-Mostafa of CalTech’s Learning from Data. Deep into the mathematical theory, Professor Abu-Mostafa begins with the question “Is learning possible?” Serious homework questions, too.

Texts: Supplementary

Programming

  • Python Data Processing

    • Seaborn: A pretty visualization library.
    • Matplotlib: Harder to use, more control.
    • Pandas: Managing related data assembled together into one “data frame”.
    • NumPy and its big cousin SciPy: numerical mathematics of all kinds.
  • Tutorials

  • Kaggle

    Use Kaggle to test the ideas that you have learned in a realistic situation. I would stick with the classic data sets for a long time; they should be challenging enough.