User Carlos Mougan

15 votes

Accepted

Why continuous features are more important than categorical features in decision tree models?

answered Jan 15, 2020 at 15:37

10 votes

Multi-country model or single model

answered Jul 20, 2020 at 9:48

10 votes

Accepted

Why SMOTE is not used in prize-winning Kaggle solutions?

answered Feb 20, 2022 at 21:35

9 votes

How can I appropriately handle cleaning of gender data?

answered Mar 20, 2020 at 9:03

8 votes

Accepted

How are samples selected from training data in Xgboost

answered Jan 8, 2020 at 9:41

7 votes

Target encoding with KFold cross-validation - how to transform test set?

answered Sep 9, 2020 at 6:17

7 votes

Accepted

How to split train/test datasets having equal classes proportion

answered Oct 11, 2020 at 15:45

6 votes

Accepted

Is there any way to explicitly measure the complexity of a Machine Learning Model in Python

answered Aug 24, 2020 at 7:56

6 votes

What is momentum in neural network?

answered Oct 18, 2020 at 9:50

6 votes

On simple 1D dataset, LogisticRegressionCV selects terrible hyperparameters, resulting scores are nonsensical

answered Apr 12, 2020 at 8:53

6 votes

Accepted

Scikit-learn OneHotEncoder effect on feature selection

answered Feb 29, 2020 at 21:46

5 votes

What is query id ("qid") in XGBoost

answered Mar 15, 2020 at 8:32

5 votes

Accepted

Are "Gradient Boosting Machines (GBM)" and GBDT exactly the same thing?

answered Sep 19, 2020 at 16:11

5 votes

Lightgbm confidence interval

answered Jun 19, 2020 at 9:15

5 votes

Accepted

One Hot Encoding for any kind of dataset

answered Jul 10, 2020 at 4:20

5 votes

Accepted

Auto ML vs Manual ML for a project

answered Jan 3, 2020 at 10:57

5 votes

Accepted

SHAP value analysis gives different feature importance on train and test set

answered Jan 24, 2020 at 17:44

5 votes

How do I encode the categorical columns if there are more than 15 unique values?

answered Dec 26, 2020 at 8:14

4 votes

How to fit a math formula to data?

answered Dec 7, 2022 at 8:40

4 votes

Extracting and Mining PDF Data

answered Jan 9, 2020 at 0:03

4 votes

Should I use keras or sklearn for PCA?

answered Jun 19, 2020 at 10:17

4 votes

Monitoring machine learning models in production

answered Oct 15, 2020 at 7:40

4 votes

Accepted

Disparity between training and testing errors with deep learning: the bias-variance tradeoff and model selection

answered Jun 1, 2020 at 7:20

3 votes

Accepted

looking for public dataset for stock market

answered Jun 13, 2020 at 13:30

3 votes

Accepted

Why 100% accuracy on test data is not good?

answered Apr 28, 2020 at 7:25

3 votes

Please review my sketch of the Machine Learning process

answered Apr 6, 2020 at 10:16

3 votes

scikit-learn RandomForestClassifier always hits 100% test accuracy

answered Apr 6, 2020 at 13:10

3 votes

How to use "tree boosting" with a data-driven loss function

answered Oct 3, 2020 at 8:36

3 votes

Accepted

Decision tree regression: Polynomials unnecessary?

answered Jan 20, 2020 at 13:20

3 votes

Document similarity

answered Jan 24, 2020 at 10:51

Stack Exchange Network

190 Answers

Why continuous features are more important than categorical features in decision tree models?

Multi-country model or single model

Why SMOTE is not used in prize-winning Kaggle solutions?

How can I appropriately handle cleaning of gender data?

How are samples selected from training data in Xgboost

Target encoding with KFold cross-validation - how to transform test set?

How to split train/test datasets having equal classes proportion

Is there any way to explicitly measure the complexity of a Machine Learning Model in Python

What is momentum in neural network?

On simple 1D dataset, LogisticRegressionCV selects terrible hyperparameters, resulting scores are nonsensical

Scikit-learn OneHotEncoder effect on feature selection

What is query id ("qid") in XGBoost

Are "Gradient Boosting Machines (GBM)" and GBDT exactly the same thing?

Lightgbm confidence interval

One Hot Encoding for any kind of dataset

Auto ML vs Manual ML for a project

SHAP value analysis gives different feature importance on train and test set

How do I encode the categorical columns if there are more than 15 unique values?

How to fit a math formula to data?

Extracting and Mining PDF Data

Should I use keras or sklearn for PCA?

Monitoring machine learning models in production

Disparity between training and testing errors with deep learning: the bias-variance tradeoff and model selection

looking for public dataset for stock market

Why 100% accuracy on test data is not good?

Please review my sketch of the Machine Learning process

scikit-learn RandomForestClassifier always hits 100% test accuracy

How to use "tree boosting" with a data-driven loss function

Decision tree regression: Polynomials unnecessary?

Document similarity