Todd MedemaTodd is a student at Carnegie Mellon University, an entrepreneur and inventor who previously cofounded AutoRef.com. You can follow his latest ventures on Twitter or on his blog.
Machine learning sounds scary - but when it's taught by Andrew Ng (of Stanford) on Coursera, it's not. Andrew does an excellent job of explaining Machine Learning in a way that's easy to understand. He focuses on the key concepts that really matter, avoiding the dangerous territory of unnecessary explanations.
The course essentially contained Supervised and Unsupervised Machine Learning with a couple algorithms in each. Ite covers a variety of learning techniques, from the simple linear and logistic regressions all the way to neural networks, clustering and recommendation engines. Each topic focuses on a practical application of the technique - these range from predicting house prices and recommending movies to controlling self-driving cars and reading hand writing. Not bad for 5 hours a week for 10 weeks!
Final grading was based off programming assignments and review questions after each section. It's the first class I've taken where you can still pass the class if you turn in everything late (you get 20% off for turning in late, and you need 80% to pass - acing the assignments and questions isn't hard, since you have unlimited attempts!)
The course is taught in Octave, a programming language specifically designed for Machine Learning that's very easy to pick up if you have any previous coding experience. Why not C++ or Python? As Andrew puts it, your time as a programmer and student is at a premium, so it's a much smarter investment to learn the concepts in Octave, and only worry about porting it to C++ if you really need the performance.While most people are terrified about Octave, it really isn't that bad. The harder part is the matrix algebra.
While his course certainly won't leave you with a PhD level knowledge, you'll leave it with plenty of confidence that you now know how to build a program that can learn from your own data sets.
Overall, I highly recommend the course as an introduction to the concepts of Machine Learning with a practical focus!
