6.867 Machine Learning -- Coursework


Data Files MATLAB Files Solutions Due Date
Homework 1 postscript PDF housing.data
postscript PDF September 20th, in class
Homework 2 postscript PDF digit_x, digit_y, digit_x_test, digit_y_test. multigaussian, gaussian2d, plotgaussians, plotgaussians3, logisticreg, sample, crossvalid. postscript PDF October 12th (in recitation)
Homework 3 postscript PDF digit_x, digit_y, digit_x_test, digit_y_test, svmexample.mat. boost_digit, boost, eval_boost, find_stump, eval_stump, trainsvm, svmpred, plotsvm2d, polyK, expK. postscript PDF November 2nd
Homework 4 postscript PDF pendigits.mat.
Raw pendigits data (not needed for PS): tared and gziped, ZIPed.
pendigits.names (not needed for PS).
Data for problem 2: tared and gziped, ZIPed.
Gaussian Mixtures code: tared and gziped, ZIPed.
Markov and HMM code: tared and gziped, ZIPed.
postscript PDF November 28th (in recitation)
Homework 5 postscript PDF - - postscript PDF do not return

The homework material is also available in the class athena locker. You can acces this by typing add 6.867 from thr athena command prompt, and then changing directory to /mit/6.867.

Collaboration Policy

Collaboration is encouraged while solving problems, however you must write your own solution to every problem. Please list your collaborators on every homework that you turn in. Copying will not be tolerated.


We will be holding one mid-term and one final.


Each student will be responsible for a project of their own selection (subject to approval). Projects will be due Wednesday, December 6th.


Your final grade will be determined from
  • Problem sets (30%)
  • Final project (25%)
  • Mid-term (15%)
  • Final (30%)
  • The percentages are approximate and may be changed at our discretion.