berndbischl coseal-algsel-benchmark-repo issues

berndbischl / coseal-algsel-benchmark-repo

6 stars 7 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Fix tests

#105 larskotthoff closed 4 years ago
0
check whether all default settings for learners are "reasonable" for results on webpage

#104 berndbischl closed 10 years ago
1
add kmeans from R for clustering

#103 berndbischl opened 10 years ago
0
Enable feature normalization only for methods if really necessary

#102 berndbischl opened 10 years ago
2
Functions should be renamed to reflect that we use "scenario" to refer to tasks now

#101 larskotthoff closed 10 years ago
4
EDA: densities. imputed values should be visible as a smaller peak

#100 berndbischl opened 10 years ago
0
R code: The whole specification of the learner in runLlamaModels is crap. Bad design

#99 berndbischl opened 10 years ago
0
R code: We should probably offer a function to subset a task, most importantly for feature steps

#98 berndbischl opened 10 years ago
0
Different imputation in EDA and Experiments

#97 kerschke closed 10 years ago
1
EDA: mark whether feature is constant or nearly constant

#96 berndbischl opened 10 years ago
0
EDA: add table or so that show on which instances all algos are equal

#95 berndbischl opened 10 years ago
1
make UCI ML data work

#94 berndbischl opened 10 years ago
0
Default feature steps must mentioned in the format

#93 berndbischl opened 10 years ago
1
Problems with names starting with number and including weird chars

#92 berndbischl opened 10 years ago
8
not features are stoch / det., but the processing steps

#91 berndbischl closed 10 years ago
1
runstatus == "ok" vs NA in algo perf

#90 berndbischl closed 10 years ago
2
shortenPathsInValidator

#89 berndbischl closed 10 years ago
2
Consistent Scenario Names

#88 kerschke closed 10 years ago
5
EDA: plotCorMatrix

#87 berndbischl closed 10 years ago
2
EDA: Algo perf summary table

#86 berndbischl closed 10 years ago
1
EDA: All plots: We CANNOT drop data

#85 berndbischl closed 10 years ago
2
findUseless seems bad

#84 berndbischl closed 10 years ago
2
Must check whether we algorithm performance values with NA, where runstatus is OK

#83 berndbischl closed 10 years ago
1
imputeCrashedRuns is bad

#82 berndbischl closed 10 years ago
11
Results: If multiple results are best they should all be colored

#81 berndbischl closed 10 years ago
2
It is completely undefined what happens if we have repetitions in the experiments

#80 berndbischl opened 10 years ago
1
EDA: potential problem with corr. clustering

#79 berndbischl closed 10 years ago
2
Random Seed vs. Repeated Experiments?

#78 mlindauer closed 10 years ago
12
Benchmark results: Mark BEST model in different red tone

#77 berndbischl closed 10 years ago
1
EDA: We should not remove crashed / timed-out algorithms in the plots

#76 berndbischl closed 10 years ago
7
addDefaultsToASTask is bad

#75 berndbischl closed 10 years ago
1
Move feature normalization function to mlr

#74 berndbischl closed 10 years ago
3
Feature imputation

#73 berndbischl closed 10 years ago
1
XMeans and classif.kknn don't work on LiDo

#72 kerschke closed 10 years ago
1
Ridge regression doesn't run without setting penalty parameter

#71 kerschke opened 10 years ago
0
validator: check NA features if corresponding step is not OK

#70 mlindauer closed 10 years ago
1
Nominal values instead of STRING in ARFF

#69 joaquinvanschoren closed 10 years ago
1
Different Algorithms on different instances

#68 kerschke closed 10 years ago
4
cluster methods: FarthestFirst SimpleKMeans maybe set nr of clusters

#67 berndbischl closed 10 years ago
30
Constant Features and Empty Classes

#66 kerschke closed 10 years ago
4
Provide orig. BBOB data

#65 berndbischl closed 10 years ago
1
llama cannot handle repeated CVs

#64 berndbischl opened 10 years ago
6
BBOB data: strange costs

#63 berndbischl opened 10 years ago
12
Potential improvements of models

#62 berndbischl opened 10 years ago
0
"presolved" solved status is misleading for instances that were solved during feature computation

#61 larskotthoff opened 10 years ago
0
Check validator

#60 kerschke opened 10 years ago
2
Check code of providing the overview table

#59 kerschke closed 10 years ago
6
finish getProvidedFeatures

#58 berndbischl closed 10 years ago
1
Why does sat11 hand / crafted have 0 feature steps?

#57 berndbischl closed 10 years ago
1
Presolved / costs

#56 berndbischl closed 10 years ago
2