User Lucas Morin - Data Science Stack Exchange

14 votes

Accepted

Why does data science see class imbalance as a problem for supervised learning when statistics does not?

answered Jan 17, 2024 at 19:50

11 votes

Rounding Float Values in ML Models

answered Apr 30, 2025 at 10:03

10 votes

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

answered Sep 1, 2023 at 19:59

8 votes

Loan prediction model relying almost entirely on Credit_History and ignoring other features

answered May 12, 2025 at 6:11

7 votes

Accepted

Duplicated features for gradient descent

answered Jan 26, 2020 at 18:14

7 votes

How to get p-value and confident interval in LogisticRegression with sklearn?

answered Mar 7, 2020 at 19:14

7 votes

Is feature engineering still useful when using XGBoost?

answered Oct 17, 2020 at 10:15

6 votes

Accepted

Is an $F_1$ score of 0.1 always bad?

answered Nov 2, 2020 at 9:35

6 votes

Why does a .fit() method change an input X to a np.array?

answered Jul 28, 2023 at 18:50

5 votes

Why are the accuracy and loss graphs of my Keras binary model not smooth?

answered Oct 4, 2024 at 11:38

5 votes

Does class weights guarantee calibration?

answered May 29, 2025 at 7:04

5 votes

ROC vs PR-score and imbalanced datasets

answered Jan 8, 2025 at 10:48

5 votes

How to adjust the hyperparameters of MLP classifier to get more perfect performance

answered Mar 3, 2020 at 12:48

5 votes

what is darknet and why is it needed for YOLO object detection?

answered Jan 6, 2020 at 10:23

4 votes

Accepted

Is it valid to compare SHAP values across models?

answered Oct 22, 2020 at 23:08

4 votes

Unsupervised Learning - Using the Outcome of Learning

answered Nov 1, 2020 at 14:02

4 votes

Time series classification, without the time dimension

answered Oct 20, 2020 at 13:50

4 votes

How to find Neural Network ZOOs?

answered Apr 8, 2025 at 7:08

4 votes

Why does prompt engineering work, since prompt engineering questions don't appear as training data?

answered Mar 13, 2024 at 14:02

3 votes

Imbalanced class in my dataset

answered Oct 29, 2024 at 15:08

3 votes

ValueError when loading sklearn DecisionTreeClassifier pickle in Python 3.10

answered Jul 27, 2024 at 6:34

3 votes

Why is it okay to set the bias vector up with zeros, and not the weight matrices?

answered Oct 21, 2020 at 21:07

3 votes

Accepted

Scikit-learn SelectKBest is picking up obviously unwanted Features

answered Oct 14, 2020 at 10:48

3 votes

How to build a model on a dataset having 40% missing values in most of the variables?

answered Oct 16, 2020 at 21:08

3 votes

Multiclass Classification with Decision Trees: Why do we calculate a score and apply softmax?

answered Mar 19, 2020 at 21:56

3 votes

Accepted

What are the model parameters in PCA?

answered Nov 14, 2020 at 8:25

3 votes

Is it possible to cluster data according to a target?

answered Dec 1, 2021 at 9:22

3 votes

Accepted

For outliers treatment: clipping, winsorizing or removing?

answered Jan 5, 2020 at 10:00

3 votes

Accepted

Sub-sampling so that sample statistics match population statistics

answered Jan 5, 2020 at 13:49

3 votes

Accepted

xgboost classifier predicted negative probabilities

answered Mar 4, 2020 at 16:47

Stack Exchange Network

170 Answers

Why does data science see class imbalance as a problem for supervised learning when statistics does not?

Rounding Float Values in ML Models

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

Loan prediction model relying almost entirely on Credit_History and ignoring other features

Duplicated features for gradient descent

How to get p-value and confident interval in LogisticRegression with sklearn?

Is feature engineering still useful when using XGBoost?

Is an $F_1$ score of 0.1 always bad?

Why does a .fit() method change an input X to a np.array?

Why are the accuracy and loss graphs of my Keras binary model not smooth?

Does class weights guarantee calibration?

ROC vs PR-score and imbalanced datasets

How to adjust the hyperparameters of MLP classifier to get more perfect performance

what is darknet and why is it needed for YOLO object detection?

Is it valid to compare SHAP values across models?

Unsupervised Learning - Using the Outcome of Learning

Time series classification, without the time dimension

How to find Neural Network ZOOs?

Why does prompt engineering work, since prompt engineering questions don't appear as training data?

Imbalanced class in my dataset

ValueError when loading sklearn DecisionTreeClassifier pickle in Python 3.10

Why is it okay to set the bias vector up with zeros, and not the weight matrices?

Scikit-learn SelectKBest is picking up obviously unwanted Features

How to build a model on a dataset having 40% missing values in most of the variables?

Multiclass Classification with Decision Trees: Why do we calculate a score and apply softmax?

What are the model parameters in PCA?

Is it possible to cluster data according to a target?

For outliers treatment: clipping, winsorizing or removing?

Sub-sampling so that sample statistics match population statistics

xgboost classifier predicted negative probabilities