User Matthew Drury

74 votes

If 900 out of 1000 people say a car is blue, what is the probability that it is blue?

answered Aug 20, 2017 at 20:29

69 votes

Accepted

Why use regularisation in polynomial regression instead of lowering the degree?

answered Jul 31, 2016 at 16:34

68 votes

Accepted

Relative variable importance for Boosting

answered Jul 29, 2015 at 3:40

61 votes

Accepted

Least Squares Regression Step-By-Step Linear Algebra Computation

answered May 28, 2015 at 22:01

55 votes

Accepted

How to apply standardization/normalization to train- and testset if prediction is the goal?

answered Sep 30, 2015 at 16:16

51 votes

Accepted

What is the difference between decision_function, predict_proba, and predict function for logistic regression problem?

answered Feb 21, 2018 at 18:01

50 votes

Accepted

One-hot vs dummy encoding in Scikit-learn

answered Jul 16, 2016 at 5:55

50 votes

Accepted

Gradient Boosting for Linear Regression - why does it not work?

answered Dec 16, 2015 at 2:18

49 votes

Accepted

What to conclude from this lasso plot (glmnet)

answered May 31, 2015 at 15:53

43 votes

Accepted

Understanding complete separation for logistic regression

answered Jul 21, 2016 at 4:37

42 votes

Accepted

Is a model fitted to data or is data fitted to a model?

answered Mar 24, 2019 at 4:44

31 votes

Accepted

What does log-uniformly distribution mean?

answered Jun 4, 2015 at 20:46

30 votes

Accepted

Does this discrete distribution have a name?

answered May 18, 2015 at 5:38

30 votes

Accepted

Why does this regression NOT fail due to perfect multicollinearity, although one variable is a linear combination of others?

answered Feb 9, 2016 at 15:58

29 votes

Cost function in OLS linear regression

answered Jun 5, 2015 at 3:22

27 votes

Accepted

Classification with Gradient Boosting : How to keep the prediction in [0,1]

answered Apr 2, 2016 at 18:14

26 votes

Are two standard normal random variables always independent?

answered Dec 7, 2016 at 15:05

24 votes

Accepted

Question about bias-variance tradeoff

answered May 18, 2015 at 21:57

23 votes

Accepted

Why are symmetric positive definite (SPD) matrices so important?

answered Jul 15, 2016 at 19:31

22 votes

What areas of mathematical statistics are highly employable?

answered Aug 26, 2016 at 22:39

21 votes

Accepted

Why linear and logistic regression coefficients cannot be estimated using same method?

answered Sep 22, 2015 at 14:53

21 votes

Accepted

Why does logistic regression produce well-calibrated models?

answered Apr 22, 2016 at 22:40

20 votes

Accepted

Why feature scaling only to training set?

answered Dec 19, 2017 at 6:33

20 votes

Is this the solution to the p-value problem?

answered Apr 1, 2016 at 5:44

20 votes

Accepted

Scikit Binomial Deviance Loss Function

answered Jun 20, 2015 at 20:43

19 votes

Accepted

Why does Central Limit Theorem break down in my simulation?

answered Mar 9, 2019 at 4:01

19 votes

Accepted

What is use of Tweedie or poisson loss/objective function in XGboost and Deep learning models

answered Oct 19, 2020 at 18:45

19 votes

When should we discretize/bin continuous independent variables/features and when should not?

answered Aug 19, 2016 at 20:37

19 votes

Should parsimony really still be the gold standard?

answered Jul 28, 2015 at 16:19

19 votes

Expected prediction error - derivation

answered Dec 5, 2015 at 23:43

Stack Exchange Network

432 Answers

If 900 out of 1000 people say a car is blue, what is the probability that it is blue?

Why use regularisation in polynomial regression instead of lowering the degree?

Relative variable importance for Boosting

Least Squares Regression Step-By-Step Linear Algebra Computation

How to apply standardization/normalization to train- and testset if prediction is the goal?

What is the difference between decision_function, predict_proba, and predict function for logistic regression problem?

One-hot vs dummy encoding in Scikit-learn

Gradient Boosting for Linear Regression - why does it not work?

What to conclude from this lasso plot (glmnet)

Understanding complete separation for logistic regression

Is a model fitted to data or is data fitted to a model?

What does log-uniformly distribution mean?

Does this discrete distribution have a name?

Why does this regression NOT fail due to perfect multicollinearity, although one variable is a linear combination of others?

Cost function in OLS linear regression

Classification with Gradient Boosting : How to keep the prediction in [0,1]

Are two standard normal random variables always independent?

Question about bias-variance tradeoff

Why are symmetric positive definite (SPD) matrices so important?

What areas of mathematical statistics are highly employable?

Why linear and logistic regression coefficients cannot be estimated using same method?

Why does logistic regression produce well-calibrated models?

Why feature scaling only to training set?

Is this the solution to the p-value problem?

Scikit Binomial Deviance Loss Function

Why does Central Limit Theorem break down in my simulation?

What is use of Tweedie or poisson loss/objective function in XGboost and Deep learning models

When should we discretize/bin continuous independent variables/features and when should not?

Should parsimony really still be the gold standard?

Expected prediction error - derivation