Giters
berndbischl
/
coseal-algsel-benchmark-repo
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
6
Watchers:
5
Issues:
104
Forks:
7
berndbischl/coseal-algsel-benchmark-repo Issues
Nominal values instead of STRING in ARFF
Closed
10 years ago
Comments count
1
It is completely undefined what happens if we have repetitions in the experiments
Updated
10 years ago
Comments count
1
runstatus == "ok" vs NA in algo perf
Closed
10 years ago
Comments count
2
not features are stoch / det., but the processing steps
Closed
10 years ago
Comments count
1
Enable feature normalization only for methods if really necessary
Updated
10 years ago
Comments count
2
llama cannot handle repeated CVs
Updated
10 years ago
Comments count
6
check whether all default settings for learners are "reasonable" for results on webpage
Closed
10 years ago
Comments count
1
Default feature steps must mentioned in the format
Updated
10 years ago
Comments count
1
add kmeans from R for clustering
Updated
10 years ago
Functions should be renamed to reflect that we use "scenario" to refer to tasks now
Closed
10 years ago
Comments count
4
EDA: potential problem with corr. clustering
Closed
10 years ago
Comments count
2
Different imputation in EDA and Experiments
Closed
10 years ago
Comments count
1
EDA: densities. imputed values should be visible as a smaller peak
Updated
10 years ago
R code: The whole specification of the learner in runLlamaModels is crap. Bad design
Updated
10 years ago
R code: We should probably offer a function to subset a task, most importantly for feature steps
Updated
10 years ago
BBOB data: strange costs
Updated
10 years ago
Comments count
12
Ridge regression doesn't run without setting penalty parameter
Updated
10 years ago
Potential improvements of models
Updated
10 years ago
EDA: add table or so that show on which instances all algos are equal
Updated
10 years ago
Comments count
1
EDA: mark whether feature is constant or nearly constant
Updated
10 years ago
XMeans and classif.kknn don't work on LiDo
Closed
10 years ago
Comments count
1
cluster methods: FarthestFirst SimpleKMeans maybe set nr of clusters
Closed
10 years ago
Comments count
30
Problems with names starting with number and including weird chars
Updated
10 years ago
Comments count
8
shortenPathsInValidator
Closed
10 years ago
Comments count
2
Consistent Scenario Names
Closed
10 years ago
Comments count
5
make UCI ML data work
Updated
10 years ago
Feature imputation
Closed
10 years ago
Comments count
1
Constant Features and Empty Classes
Closed
10 years ago
Comments count
4
Move feature normalization function to mlr
Closed
10 years ago
Comments count
3
Results: If multiple results are best they should all be colored
Closed
10 years ago
Comments count
2
addDefaultsToASTask is bad
Closed
10 years ago
Comments count
1
"presolved" solved status is misleading for instances that were solved during feature computation
Updated
10 years ago
imputeCrashedRuns is bad
Closed
10 years ago
Comments count
11
EDA: Algo perf summary table
Closed
10 years ago
Comments count
1
EDA: plotCorMatrix
Closed
10 years ago
Comments count
2
EDA: All plots: We CANNOT drop data
Closed
10 years ago
Comments count
2
findUseless seems bad
Closed
10 years ago
Comments count
2
Provide orig. BBOB data
Closed
10 years ago
Comments count
1
EDA: We should not remove crashed / timed-out algorithms in the plots
Closed
10 years ago
Comments count
7
Must check whether we algorithm performance values with NA, where runstatus is OK
Closed
10 years ago
Comments count
1
Random Seed vs. Repeated Experiments?
Closed
10 years ago
Comments count
12
Check validator
Updated
10 years ago
Comments count
2
Benchmark results: Mark BEST model in different red tone
Closed
10 years ago
Comments count
1
Different Algorithms on different instances
Closed
10 years ago
Comments count
4
validator: check NA features if corresponding step is not OK
Closed
10 years ago
Comments count
1
Add explanation of measures in experimental results
Closed
10 years ago
Comments count
12
Presolved / costs
Closed
10 years ago
Comments count
2
finish getProvidedFeatures
Closed
10 years ago
Comments count
1
Check code of providing the overview table
Closed
10 years ago
Comments count
6
Why does sat11 hand / crafted have 0 feature steps?
Closed
10 years ago
Comments count
1
Previous
Next