Lately I completed my Machine Learning course by Stanford University on Coursera. I thought it might be helpful to share my experience briefly. I chose Machine Learning because I took some courses around this topic in university years ago. Hence, I already had some (almost forgotten) knowledge and wanted to use my technical background a bit more again.
Content and Structure
The course itself is structured very well. The entire content is divided into eleven weeks, each with up to four sub-units. Amongst others, the course covers the following topics:
- Supervised learning (e.g. linear regression, logistic regression, neural networks, SVM)
- Unsupervised learning (e.g. K-means, PCA, anomaly detection)
- Fancy topics (e.g. recommender systems, OCR, MapReduce)
- Analysis & Evaluation (e.g. learning curves, error analysis, ceiling analysis)
The teacher Andrew NG presents the material in video tutorials. Precisely he explains the content using PowerPoint slides and mark-ups to write much additional information in the slides. There is no need to write anything down, because all slides including the mark-ups are available for download.
Assessments
At the end of most sub-units there are short tests with five questions each. The questions range from multiple choice over single-choice to exact result calculation. I found these tests the hardest part of the course. At the end of most weeks there is a complex programming assignment, which has to be implemented in either MATLAB or Octave (open source). I did not think the programming exercises are very complicated. The challenge was to read the instructions carefully and take only one step at a time.
Summary
The course is very time consuming. Finishing it within eleven weeks, as recommended by Coursera, requires dedicating around one to one and a half day to the course each week. Towards the end the effort becomes less. Unless you want to receive the final certificate, the course is free. I can highly recommend Coursera as a learning platform and the course Machine Learning by Andrew NG in particular.