Skip to main content
discipulus's user avatar
discipulus's user avatar
discipulus's user avatar
discipulus
  • Member for 9 years, 1 month
  • Last seen more than a week ago
  • Finland
10 votes

How exactly does Chi-square feature selection work?

9 votes

Boruta 'all-relevant' feature selection vs Random Forest 'variables of importance'

4 votes

How to do cross-validation with cv.glmnet (LASSO regression in R)?

3 votes
Accepted

Can I use oversampling with leave one out cross validation?

3 votes
Accepted

$K$-fold cross-validation: more folds or different partitioning schemes

2 votes

Is there a convincing case to be made against using last years data as a holdout set when building a predictive model?

2 votes
Accepted

How to optimize the result of K means

2 votes

Is cross validation a tool for developing performance metrics or final models?

2 votes
Accepted

Which statement in R

2 votes
Accepted

How to use cross validation for model comparison

2 votes

Variance of classification accuracy?

2 votes

How to compute F-measure and accuracy for repeated cross-validation

2 votes

What evaluation metric to use for high class imbalance where i want to capture most of the positive (ones) in the dataset

1 vote

Feature Selection Packages in R, which do both regression and classification

1 vote
Accepted

When to use training and test data sets

1 vote

What to do when the results of K-fold Cross Validation on the same dataset deviate significantly?

1 vote
Accepted

Method(s) to avoid overfitting model parameters for pre-determined model structure?

1 vote
Accepted

Cross-validation for low event rate

1 vote

How do I compute the correlation between two features and their output class?

1 vote

In k-fold cross-validation, what happens if k does not evenly divide the number of samples?

1 vote

Variable Selection One by One vs Simultaneously

1 vote
Accepted

When is n small enough to abstain from having a global test set?

0 votes

How to combine partitioned dataset to single dataset?

0 votes

randomForest::varImp VS conditional variable importance

0 votes

generating random matrix

0 votes

Does it make sense to use the top n features by importance from Random Forest in a logistic regression?

-1 votes

When people say "do analysis with and without outliers", what do they mean?