User Sam Weisenthal

14 votes

2 answers

2k views

Are Kaggle competitions just won by chance?

asked Jul 19, 2017 at 20:59

12 votes

2 answers

3k views

Why is logistic regression well calibrated, and how to ruin its calibration?

asked Feb 2, 2019 at 18:41

9 votes

1 answer

275 views

High dimensional, correlated data and top features/ covariates discovered; multiple hypothesis testing?

asked Oct 14, 2017 at 14:03

8 votes

2 answers

2k views

Apply Bayes rule sequentially

asked Jun 6, 2018 at 15:34

5 votes

3 answers

2k views

Is excluding cases with missing data fine for a predictive (not explanatory/descriptive) model?

asked Feb 23, 2018 at 23:02

5 votes

2 answers

217 views

When is $\mathbb{E}\left[\frac{1}{\sum X_{i}}\right] = \frac{1}{\mathbb{E}\left[\sum X_{i}\right]}$? [duplicate]

asked Oct 16, 2017 at 6:02

5 votes

2 answers

2k views

Is graduate level probability theory (Durett) used often in ML, DL research?

asked Aug 24, 2017 at 0:11

4 votes

1 answer

820 views

Does changing the parameter search space after nested CV introduce optimistic bias?

asked Jun 1, 2017 at 11:09

4 votes

0 answers

94 views

Predictive model for a disease based on observational data---effort in futility?

asked Jan 5, 2018 at 14:37

4 votes

1 answer

94 views

Has algorithmic fairness been studied in the past?

asked May 31, 2018 at 17:38

4 votes

2 answers

183 views

Conditional expectation $E[X\mid X=c-Y]$

asked Jul 9, 2019 at 17:28

4 votes

1 answer

202 views

If $c$ is constant, what's $E[X|X=c]$ and why?

asked Jul 9, 2019 at 21:59

4 votes

2 answers

225 views

Analysis concepts relevant to probability theory

asked Aug 30, 2018 at 22:55

3 votes

1 answer

52 views

Random variable’s distribution depends on another random variable: eg $Y\sim N(0,1/X^2)$

asked Aug 8, 2019 at 19:05

3 votes

1 answer

657 views

Why test set bootstrap in new medical prediction Nature Digital Medicine paper from Google?

asked Feb 25, 2018 at 21:36

3 votes

1 answer

84 views

Convert classifier output for disease to probability using Bayes

asked Apr 24, 2018 at 19:28

3 votes

1 answer

226 views

Maximum probability returned much less than 1: truncated calibration curve

asked Jul 1, 2017 at 23:08

3 votes

1 answer

591 views

Generative adverserial network for class imbalance

asked Jul 8, 2017 at 13:01

3 votes

1 answer

3k views

Do average coefficients in k fold cross validation resemble coefficients when trained on entire set?

asked Dec 28, 2016 at 21:32

3 votes

1 answer

828 views

If choice of learning algorithm is an unstable hyperparameter in (nested) CV, is the estimate of generalization error still valid?

asked May 27, 2017 at 23:41

3 votes

1 answer

206 views

Law of total variance and feature selection

asked Dec 14, 2017 at 1:45

3 votes

1 answer

146 views

Classification vs. Prediction

asked Jan 20, 2020 at 18:29

3 votes

1 answer

366 views

Can divergent transitions signify that I am trying to fit too complex of a model / overfitting?

asked Sep 13, 2020 at 4:39

3 votes

1 answer

812 views

What is the complement of an event in a sigma algebra?

asked Aug 30, 2018 at 0:11

2 votes

0 answers

62 views

Distribution of minimizer of sum of squared pairwise differences

asked Jun 25, 2022 at 3:19

2 votes

1 answer

480 views

Help interpreting short / truncated calibration curve

asked Jul 8, 2017 at 17:14

2 votes

2 answers

53 views

Is $P(T>c|T=k) = P(k>c|T=k)$?

asked Sep 21, 2018 at 13:35

1 vote

1 answer

107 views

Correct "hat" notation for a prediction from a GLM

asked Jan 31, 2019 at 2:37

1 vote

1 answer

53 views

If a distribution’s scale parameter cannot equal 1, is it part of a scale family?

asked Aug 16, 2019 at 0:00

1 vote

1 answer

107 views

$E[X|X=Y]=E[Y|X=Y]=E[X]=E[Y]$?

asked Jul 10, 2019 at 1:00

Stack Exchange Network

41 Questions

Are Kaggle competitions just won by chance?

Why is logistic regression well calibrated, and how to ruin its calibration?

High dimensional, correlated data and top features/ covariates discovered; multiple hypothesis testing?

Apply Bayes rule sequentially

Is excluding cases with missing data fine for a predictive (not explanatory/descriptive) model?

When is $\mathbb{E}\left[\frac{1}{\sum X_{i}}\right] = \frac{1}{\mathbb{E}\left[\sum X_{i}\right]}$? [duplicate]

Is graduate level probability theory (Durett) used often in ML, DL research?

Does changing the parameter search space after nested CV introduce optimistic bias?

Predictive model for a disease based on observational data---effort in futility?

Has algorithmic fairness been studied in the past?

Conditional expectation $E[X\mid X=c-Y]$

If $c$ is constant, what's $E[X|X=c]$ and why?

Analysis concepts relevant to probability theory

Random variable’s distribution depends on another random variable: eg $Y\sim N(0,1/X^2)$

Why test set bootstrap in new medical prediction Nature Digital Medicine paper from Google?

Convert classifier output for disease to probability using Bayes

Maximum probability returned much less than 1: truncated calibration curve

Generative adverserial network for class imbalance

Do average coefficients in k fold cross validation resemble coefficients when trained on entire set?

If choice of learning algorithm is an unstable hyperparameter in (nested) CV, is the estimate of generalization error still valid?

Law of total variance and feature selection

Classification vs. Prediction

Can divergent transitions signify that I am trying to fit too complex of a model / overfitting?

What is the complement of an event in a sigma algebra?

Distribution of minimizer of sum of squared pairwise differences

Help interpreting short / truncated calibration curve

Is $P(T>c|T=k) = P(k>c|T=k)$?

Correct "hat" notation for a prediction from a GLM

If a distribution’s scale parameter cannot equal 1, is it part of a scale family?

$E[X|X=Y]=E[Y|X=Y]=E[X]=E[Y]$?