This is the code/math I wrote in order to solve most of the assignments of Learning From Data, a machine learning course by Caltech.

Disclaimer: Don't follow this material blindly, it might be wrong. Use your judgement.

The solutions were released incrementally, one assignment at a time, after each deadline. According to the course staff, this is ok; in fact, the professor himself encourages the students to show their solutions.

To fully enjoy the course, do not look at the solutions provided here until you submit your answer.

• Octave 3.6+
• LIBSVM 3.17

Solutions for each week's homework are kept in its own directory. Each directory contains a README file that explains the main purpose of the homework along with the following subdirectories:

/img

Figures and plots generated during my thinking process

/math

Explanations and solutions involving math

/src

Octave code that answers the more practical questions

Copyright (C) Daniel Fernandes Martins. Distributed under the New BSD License.

Copyright (C) Yaser Abu-Mostafa. All rights reserved.