| Age | Commit message (Expand) | Author | Lines |
|---|
| 2024-03-12 | handle floating-point parameter values | Birte Kristina Friesel | -4/+18 |
| 2024-03-11 | CSVfile loader: get ignored and observation (attr) cols from env | Birte Kristina Friesel | -4/+40 |
| 2024-03-11 | Move Logfile and CSVfile from utils to a new loader class | Birte Kristina Friesel | -101/+107 |
| 2024-03-08 | build_rmt: handle sklearn.feature_selection.mutual_info_regression failures | Birte Kristina Friesel | -3/+9 |
| 2024-03-08 | looks like past-derf actually knew what she was doing | Birte Kristina Friesel | -0/+1 |
| 2024-03-08 | build_rmt: use DFATOOL_RMT_RELEVANCE_THRESHOLD exclusively | Birte Kristina Friesel | -9/+2 |
| 2024-03-08 | minor warning improvements | Birte Kristina Friesel | -9/+9 |
| 2024-03-08 | make FOLFunction inherit from SKLearnRegressionFunction | Birte Kristina Friesel | -16/+15 |
| 2024-03-08 | symreg hyper param | Birte Kristina Friesel | -11/+15 |
| 2024-03-08 | Add CSV support; right now hard-coded for kaggle acher linux-kernel-size | Birte Kristina Friesel | -3/+31 |
| 2024-03-08 | SKLearnRegressionFunction: export ndarray and pre-proc paramcount to dref | Birte Kristina Friesel | -45/+73 |
| 2024-03-08 | move fit_parameters check to helper function | Birte Kristina Friesel | -60/+26 |
| 2024-03-07 | feature pre-processing via information gain | Birte Kristina Friesel | -7/+112 |
| 2024-03-07 | RMT: support mutual information-based parameter relevance | Birte Kristina Friesel | -1/+28 |
| 2024-03-07 | DFATOOL_RMT_IGNORE_IRRELEVANT_PARAMS=1 → DFATOOL_RMT_RELEVANCE_METHOD=std_b... | Birte Kristina Friesel | -23/+13 |
| 2024-03-07 | Replace RMT_ENABLED=0 with MODEL=uls | Birte Kristina Friesel | -6/+6 |
| 2024-03-07 | dataref export: include environment variables | Birte Kristina Friesel | -0/+3 |
| 2024-03-07 | lightgbm: use same logger as dfatool | Birte Kristina Friesel | -0/+1 |
| 2024-03-06 | Add LightGBM support | Birte Kristina Friesel | -2/+202 |
| 2024-03-06 | boxplot_param: fix --filter-param handling | Birte Kristina Friesel | -1/+1 |
| 2024-03-06 | import lightgbm | Birte Kristina Friesel | -0/+10415 |
| 2024-02-26 | add RMT_SUBMODEL=cart | Birte Kristina Friesel | -3/+9 |
| 2024-02-26 | Logfile.load: fix errors in error output | Birte Kristina Friesel | -3/+2 |
| 2024-02-26 | add --export-csv-unparam support | Birte Kristina Friesel | -0/+36 |
| 2024-02-23 | SymbolicRegressionFunction: export hyper-parameters to dataref | Birte Kristina Friesel | -0/+13 |
| 2024-02-22 | analyze-archive: sorted model output | Birte Kristina Friesel | -4/+4 |
| 2024-02-22 | Support Symbolic Regression rather than ULS in RMT leaves | Birte Kristina Friesel | -16/+29 |
| 2024-02-22 | Replace DFATOOL_RMT_FUNCTION_LEAVES=0 with DFATOOL_RMT_SUBMODEL=static | Birte Kristina Friesel | -8/+8 |
| 2024-02-21 | SymbolicRegressor: Add hyper-parameters and complexity score calculation | Birte Kristina Friesel | -1/+78 |
| 2024-02-21 | SymReg: use MSE and set feature_names | Birte Kristina Friesel | -2/+4 |
| 2024-02-21 | black(1) | Birte Kristina Friesel | -12/+12 |
| 2024-02-21 | make SymbolicRegressionFunction inherit from SKLearnRegressionFunction | Birte Kristina Friesel | -113/+39 |
| 2024-02-21 | rename build_dtree to build_rmt | Birte Kristina Friesel | -10/+10 |
| 2024-02-21 | Rename DFATOOL_FIT_LINEAR_ONLY=1 to DFATOOL_SUBMODEL=fol | Birte Kristina Friesel | -3/+3 |
| 2024-02-21 | Rename DFATOOL_DTREE_* to DFATOOL_RMT_* | Birte Kristina Friesel | -21/+21 |
| 2024-02-21 | build_dtree: there is no need to explicitly pass param_values and data | Birte Kristina Friesel | -8/+2 |
| 2024-02-21 | Replace DFATOOL_FIT_FOL with DFATOOL_MODEL=fol | Birte Kristina Friesel | -41/+47 |
| 2024-02-21 | update documentation and examples | Birte Kristina Friesel | -20/+21 |
| 2024-02-21 | move (de)cart, lmt, xgb model generation into separate ModelAttribute functions | Birte Kristina Friesel | -84/+99 |
| 2024-02-21 | model: Remove no longer useful build_dtree wrapper | Birte Kristina Friesel | -39/+24 |
| 2024-02-21 | Always build model in get_fitted; never in constructor | Birte Kristina Friesel | -40/+23 |
| 2024-02-21 | Move CART/LMT/XGB fitting into the respective modules | Birte Kristina Friesel | -217/+270 |
| 2024-02-20 | Re-add (very very basic, for now) Symbolic Regression support | Birte Kristina Friesel | -6/+148 |
| 2024-02-19 | FOLFunction: Do not redundantly pass param_names to constructor. | Birte Kristina Friesel | -2/+1 |
| 2024-02-19 | --show-model-complexity: sort lexically | Birte Kristina Friesel | -2/+2 |
| 2024-02-19 | categorial → categorical | Birte Kristina Friesel | -48/+50 |
| 2024-02-12 | XGBoostFunction: fix complexity score and statistics calculation | Birte Kristina Friesel | -4/+7 |
| 2024-02-12 | CART, LMT export: handle function args | Birte Kristina Friesel | -2/+9 |
| 2024-02-12 | PTAModel.to_json: Handle missing states due to --filter-observations | Birte Kristina Friesel | -1/+6 |
| 2024-02-12 | unfuck param_names / feature_names handling | Birte Kristina Friesel | -169/+140 |
