summaryrefslogtreecommitdiff
path: root/lib/parameters.py
AgeCommit message (Collapse)AuthorLines
2022-02-02build_dtree: improve debug outputDaniel Friesel-7/+7
2022-02-02build_dtree: ignore codependent parametersDaniel Friesel-0/+1
2022-02-01dref export: static and analytic function ≙ single-node treeDaniel Friesel-0/+3
2022-01-31RMT generation: only split on relevant variablesDaniel Friesel-3/+19
2022-01-31ParamStats: Remove unused can_be_fitted functionDaniel Friesel-30/+0
Has been superseded by ModelAttribute.all_relevant_parameters_are_none_or_numeric
2022-01-31LMT generation: Catch LinAlg errors (e.g. during SVD), fall back to staticDaniel Friesel-1/+6
2022-01-28PTAModel: Add force_tree optionDaniel Friesel-0/+1
2022-01-27add LMT support via https://github.com/cerlymarco/linear-treeDaniel Friesel-0/+21
2022-01-26Add tree attribute export for XGBoostDaniel Friesel-2/+6
2022-01-21add depth / #nodes accessor to sklearn CART modelDaniel Friesel-2/+2
2022-01-21"if np.std(data) < 0". real galaxy brain condition right there.Daniel Friesel-1/+1
2022-01-21parameters: fix xgb model dumpDaniel Friesel-2/+3
2022-01-18optionally export xgboost modelDaniel Friesel-0/+3
2022-01-17build_tree: oops. properly set staticfunctionDaniel Friesel-2/+4
2022-01-17XGBoost: add env variables for num regressors and max depthDaniel Friesel-6/+10
2022-01-17CART, XGBoost: Fall back to StaticFunction if no parameters are availableDaniel Friesel-0/+10
2022-01-17XGBoost: Input data must be two-dimensionalDaniel Friesel-1/+1
2022-01-06param_to_ndarray: remove debug outputDaniel Friesel-3/+0
2022-01-06perform categorial→scalar conversion based on env varDaniel Friesel-2/+6
2022-01-06sklearn (CART, XGBoost): support mapping of categorial parameter values to ↵Daniel Friesel-8/+30
scalar indexes
2022-01-06add preliminary xgboost supportDaniel Friesel-0/+20
2022-01-05store decision tree attributes of xv models in dataref exportDaniel Friesel-1/+2
2021-12-23model: add sklearn CART support (CART with scalar features)Daniel Friesel-0/+48
2021-12-03dtree: switch to ssr lossDaniel Friesel-12/+13
2021-12-03optionally ignore scalar parameters during dtree split generationDaniel Friesel-14/+53
2021-12-01add ParamType helper classDaniel Friesel-0/+30
2021-11-17export number of dtree nodes to datarefDaniel Friesel-0/+4
2021-11-16add configuration variable for non-binary dtree node supportDaniel Friesel-6/+28
2021-10-25"values" is countableDaniel Friesel-1/+1
2021-10-14do not remove codependent parameters which have (None, !None) pairsDaniel Friesel-4/+19
2021-10-13codependent_param_dict: Do not throw away entirely non-numeric parameters.Daniel Friesel-2/+4
They may be important for dtree generation.
2021-08-31docsDaniel Friesel-0/+12
2021-08-26documentationDaniel Friesel-4/+22
2021-08-24dtree: Add function arg support. Update tests to reflect new models.Daniel Friesel-10/+19
2021-08-24fitting is not useful if there are less than 3 distinct int valuesDaniel Friesel-2/+12
2021-08-24support override functions in dtree codeDaniel Friesel-14/+18
2021-08-23move build_dtree method back to ModelAttribute classDaniel Friesel-0/+130
2021-08-23remove unused legacy attr.build_dtree codeDaniel Friesel-118/+2
model.build_dtree does belong into attr, but that's a different commit
2021-08-23Automatically build simple analytic or tree-based analytic model as neededDaniel Friesel-3/+21
The tree variant is used for attributes which depend on a non-numeric parameter (which can't be modeled as a function)
2021-08-19ParallelParamFit -> ParamFit, optionally without parallelismDaniel Friesel-2/+2
2021-07-12ModelAttribute, SplitFunction: Fix from_json methodsDaniel Friesel-3/+7
2021-05-10analyze-config, to_json: use paramNames in exported json nodesDaniel Friesel-2/+2
2021-04-20proper kconfig-webconf exportDaniel Friesel-0/+3
2021-03-25ParamStats: Remove unused std matrix / lut matrixDaniel Friesel-70/+15
This speeds up analysis and adds support for models with more than 32 parameters.
2021-03-17always handle co-dependent parametersDaniel Friesel-111/+80
2021-03-16move codependent parameter detection to Model / ModelAttributeDaniel Friesel-11/+62
Still TODO: Ignore codependent parameters when partitioning data for analytic modeling / regression
2021-03-16parameters: remove unused _reduce_param_matrix functionDaniel Friesel-34/+0
2021-03-15detect co-dependent parameters (wip)Daniel Friesel-1/+48
2021-03-08add dataref exportDaniel Friesel-0/+4
2021-03-05eval: remove unused function_args argumentDaniel Friesel-0/+9
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-22 23:29:25 +0000