Move to github
Forgot to add .bib file
Do I really want to add this rubbish to the dis...
Continued cleanup of doc/
Clean up of doc directory
Bug report
Bug report
Optimal agf is a bust; added decision letter fo...
I guess a negative correlation (-0.676) is bett...
not even close to working
Optimal agf should be working soon
Completed new optimal agf method, hasn't been t...
Agf optimal: optimize condition number of coeff...
Errors found using cppcheck
Finally corrected bug in borders sampling for s...
Just browsing through the code found some serio...
Documentation
Commented out debug printf statements
Direct classification embedded in multi-borders...
Incorporating direct classifiers into hierarchi...
Multi-class to interface with direct classifiers
Adding more bloat...
Minor fixes; documentation
Trying to make multi-class classification more ...
General improvements as suggested by testing wi...
Partial correlation of probabilities
Added missing logic in constrained solution (of...
Naive constraint for multi-class is a moderate ...
various...
Added directons notes; documentation
Added directons notes; documentation
Refactor more or less complete; derived multi-c...
Derived multi-class mapping works, but not well
Caught several errors
New re-factor appears to be working
Refactor of multi-borders parsing for training
Re-factoring multi-class parsing: folding it in...
Cleaning up parsing for multi-borders
Laying groundwork for externally trained multi-...
Entropy calculation probability calibration
Optimize uncertainty coefficient wrt r0
Continuum classification works, though not as w...
Removed workspace variables from multiclass cla...
Revising and debugging continuum classification
Separated out continuum retrievals to a new class
svn_clean -- something to offend the purists...
Continua-borders
Continuum retrieval using multi-borders
Use libAGF to accelerate your LIBSVM models
Version 0.9.8
Removing debug codes; passing exe ext from base...
Removed debug codes
Testing and validation
bug hunting
Documentation mostly
Gearing up for a new release
Separating out multi-borders components; (last ...
Testing multi-borders external; added experimen...
Debugging multi-borders stuff; switching to dou...
Bug hunting
various
AGF optimal actually works, but results are les...
agf optimal with multiple trials
Finally got a version of optimal agf working
Finally implemented binning method; multi class...
Random numbers resolve a tie when selecting a c...
Special multi-class classification methods whic...
Found some bugs: accelerator code seems to be w...
Fixed bug in batch_borders
Accelerator functions are a complete disaster...
Accelerator mode for multiclass training
"Accelerator" for libsvm etc.
EXE_EXT macro not set for install target in bot...
Documentation for new multi-class options
batch_borders compiled; minor errors in sample_...
Finished typing sub-routines for 'batch' border...
Starting work on 'borders accelerator'...
multi-class with external software tested on LI...
Multiclass calls external classifier in a batch...
multi_borders compatibility with other binary c...
Preprocessing routines to handle ascii files; m...
Multi-class using other binary schemes (may not...
Eliminated brain-dead binary-classifier templat...
Making multi-class easier to adapt to other bin...
Set constraint weight from command line; change...
Hacked solution to multi-class problem
Splitting up multi-class classes; adding more o...
Hacked together multi-class scheme that combine...
multi-class: made sum of cond. p=1 hard constra...
Constrained solution for multi-class conditiona...
Evaluation routines
Improving evaluation functions
Improving evaluation functions--new module
Cleanup: deleted program that separates multi-c...
Added size ratio parameter in sc_borders
Converting #define's into const's
New interface for supernewton
why the fuck do I keep working on this shit?