index
:
dfatool
decisiontrees
feature/lennart/get_files_array
kconfig-modeling
lennart
main
master
merge-prep/janis
merge-prep/lennart
state-duration-timers
Performance Model Generation for Configurable System Components
derf
summary
refs
log
tree
commit
diff
log msg
author
committer
range
path:
root
/
lib
Age
Commit message (
Collapse
)
Author
Lines
2022-01-27
add LMT support via https://github.com/cerlymarco/linear-tree
Daniel Friesel
-0
/
+2961
2022-01-26
Add tree attribute export for XGBoost
Daniel Friesel
-10
/
+32
2022-01-21
add depth / #nodes accessor to sklearn CART model
Daniel Friesel
-2
/
+28
2022-01-21
"if np.std(data) < 0". real galaxy brain condition right there.
Daniel Friesel
-1
/
+1
2022-01-21
parameters: fix xgb model dump
Daniel Friesel
-2
/
+3
2022-01-18
optionally export xgboost model
Daniel Friesel
-0
/
+3
2022-01-18
loader: gracefully handle uncreatable cache directory
Daniel Friesel
-0
/
+6
2022-01-17
build_tree: oops. properly set staticfunction
Daniel Friesel
-2
/
+4
2022-01-17
XGBoost: add env variables for num regressors and max depth
Daniel Friesel
-6
/
+10
2022-01-17
CART, XGBoost: Fall back to StaticFunction if no parameters are available
Daniel Friesel
-0
/
+10
2022-01-17
XGBoost: Input data must be two-dimensional
Daniel Friesel
-1
/
+1
2022-01-14
fix categorial-to-scalar handler for categorial values unseen during training
Daniel Friesel
-1
/
+1
2022-01-10
SKLearnRegressionFunction: handle unseen params in categorial_to_index mode
Daniel Friesel
-2
/
+13
2022-01-06
param_to_ndarray: remove debug output
Daniel Friesel
-3
/
+0
2022-01-06
perform categorial→scalar conversion based on env var
Daniel Friesel
-2
/
+6
2022-01-06
sklearn (CART, XGBoost): support mapping of categorial parameter values to ↵
Daniel Friesel
-10
/
+36
scalar indexes
2022-01-06
add preliminary xgboost support
Daniel Friesel
-1
/
+27
2022-01-06
add xv data to dref export
Daniel Friesel
-1
/
+1
2022-01-05
store decision tree attributes of xv models in dataref export
Daniel Friesel
-3
/
+27
2022-01-05
cross validation: return intermediate models used for XV
Daniel Friesel
-3
/
+5
These are interesting for statistics, e.g. to determine the average dtree size
2021-12-23
model: add sklearn CART support (CART with scalar features)
Daniel Friesel
-3
/
+91
2021-12-03
dtree: switch to ssr loss
Daniel Friesel
-12
/
+13
2021-12-03
model: fix parameter counts in dataref export
Daniel Friesel
-1
/
+3
2021-12-03
model: Remove temporary ResKIL accuracy/power modeling hack
Daniel Friesel
-7
/
+0
2021-12-03
optionally ignore scalar parameters during dtree split generation
Daniel Friesel
-16
/
+67
2021-12-01
add ParamType helper class
Daniel Friesel
-13
/
+48
2021-12-01
model: Add MAPE export
Daniel Friesel
-0
/
+12
2021-12-01
move some standard arguments to cli helper class
Daniel Friesel
-5
/
+22
2021-11-24
validation: add SMAPE output
Daniel Friesel
-2
/
+3
2021-11-23
add parameter-aware cross validation
Daniel Friesel
-3
/
+40
2021-11-22
handle fit errors in cross validation
Daniel Friesel
-4
/
+26
2021-11-19
kconfig exploration: 5 values should be sufficient
Daniel Friesel
-1
/
+1
2021-11-17
export number of dtree nodes to dataref
Daniel Friesel
-0
/
+13
2021-11-16
add configuration variable for non-binary dtree node support
Daniel Friesel
-13
/
+56
2021-11-16
Add env var for safe functions; disable unsafe functions in that case
Daniel Friesel
-1
/
+7
2021-11-12
dataref: export number of samples per measurement class
Daniel Friesel
-0
/
+2
2021-11-11
AnalyticModel dref export: Add number of boolean/scalar/enum parameters
Daniel Friesel
-1
/
+31
2021-11-03
Set DFATOOL_FIT_LINEAR_ONLY=1 to omit non-linear functions from regression
Daniel Friesel
-18
/
+22
2021-11-01
add static & param model output to analyze-kconfig
Daniel Friesel
-0
/
+22
2021-11-01
move model quality table to dfatool.cli; add quality output to analyze-kconfig
Daniel Friesel
-0
/
+51
2021-10-28
kconfig loader: do not use frozendicts (json doesn't know about them)
Daniel Friesel
-3
/
+2
2021-10-28
analyze-kconfig: add dataref export
Daniel Friesel
-18
/
+25
2021-10-27
kconfig: add DFATOOL_KCONF_IGNORE_NUMERIC / _STRING options
Daniel Friesel
-1
/
+10
2021-10-26
allow custom standard deviation thresholds for decision tree compilation
Daniel Friesel
-4
/
+26
2021-10-26
kconfig loader: load kconfig from its base directory
Daniel Friesel
-13
/
+18
this ensures that source statements are handled correctly
2021-10-26
kconfig: enter project directory before loading kconfig
Daniel Friesel
-3
/
+18
Otherwise, "source" statements will break
2021-10-26
kconfig: always store config hash; handle randconfig without KCONFIG_SEED
Daniel Friesel
-1
/
+4
2021-10-25
"values" is countable
Daniel Friesel
-1
/
+1
2021-10-25
dataref export: handle missing MAPE entries
Daniel Friesel
-9
/
+21
2021-10-25
model: Remove legacy with_function_leaves override
Daniel Friesel
-5
/
+6
[prev]
[next]