1
1
Machine Learning Class Projects
Greg Grudic
2
General Info
• You must choose a project by next week!
• The project is due on December 9.
Submission via email:
– 8 page write-up in NIPS format (see nips.cc)
– Source code used (if any)
3
Density Estimation Algorithm in
Homework 4
• Implement
– Sparse models
– Alpha optimization
– Extension to hundreds of thousands of training
examples
• Test on Standard datasets and compare to
published results
4
Mobile Robot Path Following in
Outdoor Environments: Project 1
• Apply many algorithms to more data of the type
used in Homework 4.
• Use the WEKA toolkit and perhaps LIBSVM
• Report on
– Which gives the best error rates
– Which model is the fastest to evaluate
• Close the loop with the best model on the real
robot (limited test – I just want you to take a
model and put it into the loop).
