User Sam Weisenthal

36 votes

Accepted

What is the intuition behind the formula for conditional probability?

answered Feb 1, 2018 at 13:48

25 votes

Scikit correct way to calibrate classifiers with CalibratedClassifierCV

answered Jul 2, 2017 at 19:40

21 votes

Predictive models: statistics can't possibly beat machine learning?

answered Mar 24, 2018 at 12:14

18 votes

Accepted

Why is data augmentation classified as a type of regularization?

answered Jul 31, 2017 at 13:00

17 votes

Accepted

AUPRC vs. AUC-ROC?

answered Apr 5, 2018 at 19:30

15 votes

Accepted

Adjusting probability threshold for sklearn's logistic regression model

answered Dec 31, 2018 at 1:42

14 votes

Who invented stochastic gradient descent?

answered Nov 14, 2017 at 14:14

12 votes

High Recall - Low Precision for unbalanced dataset

answered Jul 9, 2017 at 20:00

11 votes

Why is $SST=SSE + SSR$? (One variable linear regression)

answered Aug 2, 2019 at 19:31

10 votes

machine learning techniques for longitudinal data

answered Jan 20, 2017 at 14:34

8 votes

Accepted

Confusion in derivation of bias variance decomposition

answered Feb 2, 2018 at 23:37

8 votes

Perform cross-validation on train set or entire data set

answered Apr 10, 2018 at 19:13

8 votes

What is the "binary:logistic" objective function in XGBoost?

answered Apr 24, 2018 at 19:42

8 votes

In binary classification, in what specific case should I use accuracy, AUROC, logloss or Brier score to choose model?

answered May 27, 2018 at 21:46

7 votes

Accepted

How to split dataset for model selection and tuning

answered Jul 13, 2017 at 14:00

6 votes

Accepted

Deep Learning: Wild differences after model is retrained on the same data, what to do?

answered Aug 2, 2017 at 19:04

6 votes

Accepted

Time-series logistic regression

answered Jul 3, 2017 at 12:13

6 votes

Is it possible to use Recurrent NN (LSTM) for classification?

answered Nov 12, 2017 at 19:48

5 votes

Deriving gradient of a single layer neural network w.r.t its inputs, what is the operator in the chain rule?

answered Aug 3, 2017 at 5:05

4 votes

Accepted

How does increasing the dataset decrease the cross validation error in case of large variance?

answered Jul 21, 2017 at 18:58

4 votes

How to speed up training of a Neural Network?

answered Sep 8, 2017 at 17:53

4 votes

Posterior distribution and MCMC

answered May 3, 2018 at 12:41

4 votes

Accepted

In spite of the condition $i \ne j$, why is $E[Y_iY_j] = \mu^2$?

answered May 21, 2019 at 20:33

3 votes

Can the mean squared error be used for classification?

answered Jan 5, 2020 at 17:40

3 votes

What will a neural network respond to a never-seen feature?

answered Jun 15, 2018 at 17:31

3 votes

Accepted

Classification for high dimension data

answered Mar 9, 2018 at 3:50

3 votes

Ergodicity explained in layman terms

answered May 7, 2018 at 16:16

3 votes

Accepted

Generative adverserial network for class imbalance

answered Sep 3, 2017 at 19:34

3 votes

Imputing missing values on a testing set

answered Sep 4, 2017 at 15:17

3 votes

How to model longitudinal big data?

answered Jan 20, 2017 at 14:41

Stack Exchange Network

124 Answers

What is the intuition behind the formula for conditional probability?

Scikit correct way to calibrate classifiers with CalibratedClassifierCV

Predictive models: statistics can't possibly beat machine learning?

Why is data augmentation classified as a type of regularization?

AUPRC vs. AUC-ROC?

Adjusting probability threshold for sklearn's logistic regression model

Who invented stochastic gradient descent?

High Recall - Low Precision for unbalanced dataset

Why is $SST=SSE + SSR$? (One variable linear regression)

machine learning techniques for longitudinal data

Confusion in derivation of bias variance decomposition

Perform cross-validation on train set or entire data set

What is the "binary:logistic" objective function in XGBoost?

In binary classification, in what specific case should I use accuracy, AUROC, logloss or Brier score to choose model?

How to split dataset for model selection and tuning

Deep Learning: Wild differences after model is retrained on the same data, what to do?

Time-series logistic regression

Is it possible to use Recurrent NN (LSTM) for classification?

Deriving gradient of a single layer neural network w.r.t its inputs, what is the operator in the chain rule?

How does increasing the dataset decrease the cross validation error in case of large variance?

How to speed up training of a Neural Network?

Posterior distribution and MCMC

In spite of the condition $i \ne j$, why is $E[Y_iY_j] = \mu^2$?

Can the mean squared error be used for classification?

What will a neural network respond to a never-seen feature?

Classification for high dimension data

Ergodicity explained in layman terms

Generative adverserial network for class imbalance

Imputing missing values on a testing set

How to model longitudinal big data?