| Age | Commit message (Collapse) | Author | Files | Lines | |
|---|---|---|---|---|---|
| 2016-12-07 | allow datasets to be stored in libsvm/svmlight format | Gertjan van den Burg | 14 | -19/+783 | |
| 2016-12-07 | moved check for class labels to seperate module | Gertjan van den Burg | 8 | -44/+286 | |
| 2016-12-07 | throw warning when using sparse matrices with kernels | Gertjan van den Burg | 2 | -0/+31 | |
| 2016-12-07 | make test for sparsity the same everywhere | Gertjan van den Burg | 1 | -4/+3 | |
| 2016-12-07 | only print kernel computation method with nonlinear kernels | Gertjan van den Burg | 1 | -2/+4 | |
| 2016-12-07 | print warning when a test dataset is supplied in training file | Gertjan van den Burg | 1 | -2/+6 | |
| 2016-12-07 | merge types into globals | Gertjan van den Burg | 2 | -52/+26 | |
| 2016-12-07 | add ifndef and GENSVM_ before defines | Gertjan van den Burg | 13 | -97/+105 | |
| 2016-12-07 | switch some integer variables to long for cross platform size guarantees | Gertjan van den Burg | 10 | -21/+21 | |
| 2016-12-06 | document undocumented elements | Gertjan van den Burg | 16 | -86/+346 | |
| 2016-12-06 | add msvmmaj matlab file to git | Gertjan van den Burg | 1 | -0/+361 | |
| 2016-12-06 | add unit tests for gensvm_train | Gertjan van den Burg | 5 | -3/+530 | |
| 2016-12-05 | fix missing test in gensvm_optimize and expose doublesort | Gertjan van den Burg | 6 | -22/+74 | |
| 2016-12-05 | add octave testfiles to git | Gertjan van den Burg | 11 | -0/+809 | |
| 2016-12-05 | further unit tests for kernel module | Gertjan van den Burg | 5 | -39/+3804 | |
| 2016-12-05 | Tests and documentation kernel module | Gertjan van den Burg | 4 | -92/+334 | |
| 2016-12-05 | Improvements and tests for tasks and gridsearch | Gertjan van den Burg | 6 | -15/+1426 | |
| 2016-11-07 | compute ZAZ in blocks to increase numerical precision | Gertjan van den Burg | 3 | -23/+64 | |
| 2016-11-03 | prepare for gridsearch unit testing | Gertjan van den Burg | 15 | -394/+867 | |
| 2016-10-24 | update copyright information | Gertjan van den Burg | 73 | -492/+2045 | |
| 2016-10-18 | fix problems with dense ZV computation when data->n < model->n | Gertjan van den Burg | 2 | -5/+142 | |
| 2016-10-17 | minor documentation fixes | Gertjan van den Burg | 6 | -22/+11 | |
| 2016-10-17 | refactor gensvm_pred to gensvm_predict | Gertjan van den Burg | 5 | -6/+6 | |
| 2016-10-17 | Update predictions to work with sparse matrices | Gertjan van den Burg | 8 | -61/+508 | |
| This is done by pulling the Z*V routines from the gensvm_optimize file to a seperate file, since they are shared by prediction and get_loss | |||||
| 2016-10-17 | update doxyfile | Gertjan van den Burg | 1 | -126/+234 | |
| 2016-10-17 | Add functionality for cv_util for sparse matrices | Gertjan van den Burg | 3 | -7/+342 | |
| 2016-10-17 | Create debug function for printing GenSparse structs | Gertjan van den Burg | 5 | -6/+106 | |
| 2016-10-14 | remove temporary test seed setting | Gertjan van den Burg | 1 | -4/+1 | |
| 2016-10-14 | documentation fixes | Gertjan van den Burg | 18 | -42/+56 | |
| 2016-10-14 | add sparse matrices to GenSVM and reorganize update functionality | Gertjan van den Burg | 18 | -1276/+2164 | |
| 2016-10-06 | documentation fixes | Gertjan van den Burg | 5 | -3/+28 | |
| 2016-10-06 | make use of the fact that the first element of z_i is always 1 | Gertjan van den Burg | 1 | -3/+8 | |
| 2016-10-06 | change timer functionality to use clock_monotonic_raw | Gertjan van den Burg | 4 | -19/+76 | |
| 2016-10-06 | Coding style - set all pointers to NULL initially and after free | Gertjan van den Burg | 10 | -25/+58 | |
| 2016-10-06 | Switch to using dsyrk instead of dsyr for speed. | Gertjan van den Burg | 6 | -65/+282 | |
| Also added a workspace (GenWork) structure for to hold working matrices for the gensvm_get_update() and gensvm_get_loss() functions | |||||
| 2016-10-05 | documentation fixes | Gertjan van den Burg | 2 | -8/+3 | |
| 2016-09-30 | rewrite of the update function to use only rank 1 operations where possible | Gertjan van den Burg | 4 | -162/+110 | |
| 2016-09-30 | Rewrite UU matrix to be K*K*(K-1) instead of n*K*(K-1) | Gertjan van den Burg | 8 | -281/+236 | |
| significant memory reduction by turning the 3D UU matrix into a 2D block matrix, with significantly less dimensions | |||||
| 2016-09-30 | Remove category matrix from implementation | Gertjan van den Burg | 5 | -120/+24 | |
| It's a remnant from the Matlab implementation, where it actually makes sense to use. Here it becomes a significant memory burden for large datasets, and doesn't provide any speedup | |||||
| 2016-09-30 | remove unnecessary W and t matrices | Gertjan van den Burg | 3 | -24/+2 | |
| 2016-09-30 | minor changes and start on test of gensvm_optimize | Gertjan van den Burg | 4 | -8/+120 | |
| 2016-09-30 | use atlas for optimized blas/lapack | Gertjan van den Burg | 1 | -1/+1 | |
| 2016-09-30 | Add define for print_iter | Gertjan van den Burg | 1 | -1/+6 | |
| 2016-09-30 | Break up update function and add unit tests for parts | Gertjan van den Burg | 3 | -231/+751 | |
| 2016-09-30 | fix typo in file io | Gertjan van den Burg | 1 | -1/+1 | |
| 2016-09-30 | reorder initialization and allocation | Gertjan van den Burg | 1 | -3/+3 | |
| 2016-09-30 | Prevent memory leaks when copying data filename | Gertjan van den Burg | 1 | -1/+2 | |
| 2016-09-30 | Exclude test folders for documentation | Gertjan van den Burg | 1 | -1/+1 | |
| 2016-09-26 | moved mainpage to README | Gertjan van den Burg | 2 | -29/+20 | |
| 2016-09-22 | More unit tests for optimize | Gertjan van den Burg | 5 | -6/+876 | |
| Also added auxiliary Octave files used to construct the unit tests | |||||
