The machine learning course at Coursera is a very good introduction to machine learning. It presents a systematic method for developing machine learning applications. The method can help you avoid time consuming tasks like collecting more data for your application if in fact the effort should be put in another part of your application. Based on experience from industry applications developed in Silicon Valley the method can be of help for experienced developers of machine learning applications as well.
Online course in machine learning.
The following lists some features of the course:
- Systematic development method for machine learning applications.
- Introduction to GNU Octave. This is an open source programming language quite similar to Matlab. Octave helps you do scientific computations by enabling you to express things like matrix computations in a clear and concise manner. Read more about Octave.
- Examples of industrial machine learning applications like digit recognition and autonomous car driving.
- Machine learning concepts like supervised and unsupervised learning.
- Machine learning algorithms like linear regression, logistic regression, neural networks, support vector machines and more.
- Video lectures enabling you to stop and play parts over again if there is something you need to have repeated.
- Review exercises to check your understanding of the stuff presented in the video lectures.
- Programming exercises to solidify your knowledge of the stuff presented in the video lectures and to grade you.
- Ideas for getting started in machine learning. If you have no idea of what to do in order to get started in developing machine learning applications, then this course might help.
- Good advice to prevent wasting time. Cases are mentioned where groups have
spent half a year on collecting data for their machine learning application just to find out that
adding more data did not help. The more systematic development method presented in the course might
have helped them by showing that collecting more data would not increase performance.
Advice are given for both single algorithms and for applications consisting of many machine learning components.
- Good and pedagogical teacher.
- Discussion forums.
Visit the course web site here.