User Frank Harrell

127 votes

Accepted

If mean is so sensitive, why use it in the first place?

answered Aug 13, 2011 at 16:46

69 votes

Given the power of computers these days, is there ever a reason to do a chi-squared test rather than Fisher's exact test?

chi-squared-test contingency-tables fishers-exact-test

answered Aug 14, 2011 at 0:29

64 votes

Accepted

Including the interaction but not the main effects in a model

regression modeling interaction regression-coefficients

answered May 21, 2011 at 12:31

58 votes

T-test for non normal when N>50?

hypothesis-testing normal-distribution t-test inference central-limit-theorem

answered Apr 20, 2011 at 12:59

57 votes

What does AUC stand for and what is it?

classification prediction roc auc abbreviation

answered Jan 31, 2016 at 14:32

49 votes

Skills hard to find in machine learners?

machine-learning data-mining

answered Jun 24, 2014 at 11:29

42 votes

Accepted

If only prediction is of interest, why use lasso over ridge?

machine-learning prediction lasso regularization ridge-regression

answered Mar 5, 2018 at 13:28

41 votes

Accepted

Why is the use of high order polynomials for regression discouraged?

regression splines polynomial interpolation

answered Oct 20, 2021 at 12:34

35 votes

The Two Cultures: statistics vs. machine learning?

machine-learning pac-learning

answered Mar 18, 2015 at 12:14

35 votes

Is normality testing 'essentially useless'?

hypothesis-testing normality-assumption philosophical

answered Aug 1, 2013 at 11:52

32 votes

Accepted

Why is AUC higher for a classifier that is less accurate than for one that is more accurate?

machine-learning classification roc auc bayesian-network

answered Mar 20, 2014 at 12:35

31 votes

Accepted

How to draw valid conclusions from "big data"?

data-mining dataset large-data validation

answered Feb 9, 2012 at 13:47

30 votes

What's wrong with Bonferroni adjustments?

hypothesis-testing multiple-comparisons bonferroni

answered Oct 17, 2014 at 13:32

30 votes

How to choose between t-test or non-parametric test e.g. Wilcoxon in small samples

hypothesis-testing t-test nonparametric small-sample wilcoxon-mann-whitney-test

answered Nov 4, 2014 at 12:48

29 votes

Should covariates that are not statistically significant be 'kept in' when creating a model?

regression statistical-significance model ancova faq

answered Aug 3, 2013 at 19:20

29 votes

Accepted

When is logistic regression suitable?

regression machine-learning logistic classification regression-strategies

answered Feb 23, 2016 at 12:32

29 votes

Accepted

Evaluating logistic regression and interpretation of Hosmer-Lemeshow Goodness of Fit

r logistic goodness-of-fit regression-strategies model-evaluation

answered Apr 15, 2016 at 12:37

28 votes

When will L1 regularization work better than L2 and vice versa?

regression lasso regularization ridge-regression

answered Nov 28, 2015 at 17:15

27 votes

What happens if the explanatory and response variables are sorted independently before regression?

regression correlation

answered Dec 9, 2015 at 0:32

25 votes

Diagnostics for logistic regression?

regression logistic diagnostic

answered Sep 20, 2013 at 11:36

24 votes

Accepted

How to translate the output from an lm() fit with a cubic spline into a regression equation

r splines

answered Aug 16, 2013 at 20:30

24 votes

Accepted

Can the MIC algorithm for detecting non-linear correlations be explained intuitively?

correlation nonparametric bioinformatics information-theory mutual-information

answered Dec 23, 2011 at 16:00

24 votes

Is it valid to include a baseline measure as control variable when testing the effect of an independent variable on change scores?

regression repeated-measures least-squares change-scores

answered Sep 20, 2011 at 11:45

24 votes

Logistic Regression - Error Term and its Distribution

logistic binomial-distribution bernoulli-distribution logistic-distribution

answered Nov 20, 2014 at 12:42

24 votes

Accepted

Model building and selection using Hosmer et al. 2013. Applied Logistic Regression in R

r logistic model-selection regression-strategies

answered Oct 7, 2014 at 21:54

24 votes

ASA discusses limitations of $p$-values - what are the alternatives?

hypothesis-testing bayesian p-value frequentist

answered Mar 11, 2016 at 17:05

23 votes

What is the difference between test set and validation set?

machine-learning validation

answered Dec 15, 2014 at 3:39

23 votes

How do I test that two continuous variables are independent?

hypothesis-testing references independence

answered Jun 16, 2014 at 19:44

22 votes

Is there any reason to prefer the AIC or BIC over the other?

modeling aic cross-validation bic model-selection

answered Apr 30, 2011 at 2:01

22 votes

Why don't we use absolute values of differences instead of squared differences in the denominator of the correlation coefficient?

correlation

answered Aug 28, 2021 at 11:40

Stack Exchange Network

current community

your communities

more stack exchange communities

1,353 Answers

If mean is so sensitive, why use it in the first place?

Given the power of computers these days, is there ever a reason to do a chi-squared test rather than Fisher's exact test?

Including the interaction but not the main effects in a model

T-test for non normal when N>50?

What does AUC stand for and what is it?

Skills hard to find in machine learners?

If only prediction is of interest, why use lasso over ridge?

Why is the use of high order polynomials for regression discouraged?

The Two Cultures: statistics vs. machine learning?

Is normality testing 'essentially useless'?

Why is AUC higher for a classifier that is less accurate than for one that is more accurate?

How to draw valid conclusions from "big data"?

What's wrong with Bonferroni adjustments?

How to choose between t-test or non-parametric test e.g. Wilcoxon in small samples

Should covariates that are not statistically significant be 'kept in' when creating a model?

When is logistic regression suitable?

Evaluating logistic regression and interpretation of Hosmer-Lemeshow Goodness of Fit

When will L1 regularization work better than L2 and vice versa?

What happens if the explanatory and response variables are sorted independently before regression?

Diagnostics for logistic regression?

How to translate the output from an lm() fit with a cubic spline into a regression equation

Can the MIC algorithm for detecting non-linear correlations be explained intuitively?

Is it valid to include a baseline measure as control variable when testing the effect of an independent variable on change scores?

Logistic Regression - Error Term and its Distribution

Model building and selection using Hosmer et al. 2013. Applied Logistic Regression in R

ASA discusses limitations of $p$-values - what are the alternatives?

What is the difference between test set and validation set?

How do I test that two continuous variables are independent?

Is there any reason to prefer the AIC or BIC over the other?

Why don't we use absolute values of differences instead of squared differences in the denominator of the correlation coefficient?