User IcannotFixThis

173 votes

2 answers

208k views

A list of cost functions used in neural networks, alongside applications

machine-learning neural-networks

Phylliida

2,835

asked May 31, 2015 at 19:37

131 votes

6 answers

71k views

Why are neural networks becoming deeper, but not wider?

Karnivaurus

6,209

asked Jul 9, 2016 at 6:35

93 votes

8 answers

31k views

When is unbalanced data really a problem in Machine Learning?

machine-learning classification predictive-models unbalanced-classes faq

Tim♦

117k

asked Jun 2, 2017 at 12:08

59 votes

5 answers

29k views

What are disadvantages of state-space models and Kalman Filter for time-series modelling?

time-series arima kalman-filter vector-autoregression

Kochede

2,047

asked Dec 2, 2013 at 10:53

44 votes

1 answer

20k views

What is the difference between conditional and unconditional quantile regression?

quantile-regression

AlexH

946

asked Dec 22, 2013 at 12:39

40 votes

7 answers

8k views

Why is the null hypothesis often sought to be rejected?

hypothesis-testing

Prometheus

906

asked Dec 4, 2015 at 14:44

36 votes

2 answers

4k views

Understanding p-value

hypothesis-testing p-value interpretation

JDL

441

asked Nov 30, 2012 at 13:14

26 votes

2 answers

15k views

R-squared in quantile regression

r-squared quantile-regression

rnso

9,193

asked Dec 15, 2014 at 17:21

25 votes

2 answers

3k views

Restricted Boltzmann Machine : how is it used in machine learning?

machine-learning neural-networks deep-learning deep-belief-networks restricted-boltzmann-machine

Bill Ancalagon the black

1,744

asked Jun 28, 2016 at 0:41

21 votes

2 answers

741 views

Fisher information in a hierarchical model

multilevel-analysis fisher-information

emakalic

2,009

asked Mar 28, 2011 at 6:33

17 votes

4 answers

10k views

How meaningful is the connection between MLE and cross entropy in deep learning?

maximum-likelihood deep-learning cross-entropy

orome

378

asked Aug 13, 2017 at 19:03

15 votes

1 answer

1k views

Why don't we use non-constant learning rates for gradient decent for things other then neural networks?

machine-learning deep-learning optimization gradient-descent

Tim♦

117k

asked Mar 2, 2018 at 14:47

15 votes

1 answer

854 views

Pooling calibration plots after multiple imputation

data-visualization data-imputation multiple-imputation pooling calibration

IWS

2,594

asked Mar 7, 2017 at 15:30

14 votes

1 answer

4k views

AIC of ridge regression: degrees of freedom vs. number of parameters

regression aic ridge-regression degrees-of-freedom

Funkwecker

2,552

asked Dec 2, 2016 at 11:34

13 votes

1 answer

4k views

Calculating confidence intervals via bootstrap on dependent observations

confidence-interval bootstrap hidden-markov-model

Sadeghd

413

asked Aug 13, 2011 at 11:28

11 votes

3 answers

2k views

Is autocorrelation in a supervised learning dataset a problem?

time-series random-forest autocorrelation supervised-learning nonlinear

rinspy

3,248

asked Jul 26, 2017 at 9:44

11 votes

1 answer

1k views

A paper mentions a "Monte Carlo simulation to determine the number of principal components"; how does it work?

pca eigenvalues neuroimaging parallel-analysis

chainhomelow

359

asked Jan 6, 2016 at 1:25

10 votes

2 answers

3k views

Can I use bootstrapping to estimate the uncertainty in a maximum value of a GAM?

bootstrap monte-carlo generalized-additive-model uncertainty

Jan-Erik Thrane

103

asked Jan 12, 2016 at 15:55

9 votes

0 answers

615 views

Linear regression with overlapping observations

time-series least-squares linear-model

rinspy

3,248

asked Aug 15, 2017 at 8:44

8 votes

2 answers

2k views

Correlated cases and Cross Validation

machine-learning classification cross-validation autocorrelation

Edu

531

asked Feb 9, 2015 at 22:24

7 votes

4 answers

2k views

Gradient boosting - extreme predictions vs predictions close to 0.5

classification cart boosting

rinspy

3,248

asked May 9, 2017 at 8:15

7 votes

1 answer

2k views

Why deep learning prefer the probability distribution with a sharp point?

probability distributions deep-learning exponential-distribution laplace-distribution

Flamingo

191

asked Apr 25, 2017 at 15:15

7 votes

2 answers

2k views

Why is the restricted boltzmann machine both unsupervised and generative?

deep-learning unsupervised-learning restricted-boltzmann-machine

user46925

asked Aug 5, 2014 at 12:38

7 votes

1 answer

2k views

Comparing approaches of MLE estimates of a Weibull distribution

maximum-likelihood fitting weibull-distribution

philip

113

asked Aug 16, 2016 at 14:19

6 votes

3 answers

2k views

What problem does Residual Nets solve that batch normalization does not solve?

machine-learning neural-networks deep-learning

Charlie Parker

6,096

asked May 6, 2017 at 15:58

6 votes

1 answer

2k views

Energy function of RBM

machine-learning markov-process

mchen

710

asked Jun 11, 2013 at 11:55

5 votes

2 answers

1k views

Factors with only two variables in factor analysis

factor-analysis eigenvalues

user44308

51

asked Apr 23, 2014 at 9:43

5 votes

2 answers

2k views

Approximate mean from .25, .5, .75 percentiles

mean descriptive-statistics median

IcannotFixThis

1,162

asked Apr 13, 2015 at 7:12

5 votes

2 answers

1k views

Exercises in EM with censored data

self-study maximum-likelihood expectation-maximization conditional-expectation exponential-family

Jon

2,230

asked May 13, 2017 at 19:24

5 votes

1 answer

8k views

How many AdaBoost iterations?

convergence boosting adaboost

shadowtalker

11.6k

asked Jul 28, 2015 at 13:56

Stack Exchange Network

current community

your communities

more stack exchange communities

67 Bookmarks

A list of cost functions used in neural networks, alongside applications

Why are neural networks becoming deeper, but not wider?

When is unbalanced data really a problem in Machine Learning?

What are disadvantages of state-space models and Kalman Filter for time-series modelling?

What is the difference between conditional and unconditional quantile regression?

Why is the null hypothesis often sought to be rejected?

Understanding p-value

R-squared in quantile regression

Restricted Boltzmann Machine : how is it used in machine learning?

Fisher information in a hierarchical model

How meaningful is the connection between MLE and cross entropy in deep learning?

Why don't we use non-constant learning rates for gradient decent for things other then neural networks?

Pooling calibration plots after multiple imputation

AIC of ridge regression: degrees of freedom vs. number of parameters

Calculating confidence intervals via bootstrap on dependent observations

Is autocorrelation in a supervised learning dataset a problem?

A paper mentions a "Monte Carlo simulation to determine the number of principal components"; how does it work?

Can I use bootstrapping to estimate the uncertainty in a maximum value of a GAM?

Linear regression with overlapping observations

Correlated cases and Cross Validation

Gradient boosting - extreme predictions vs predictions close to 0.5

Why deep learning prefer the probability distribution with a sharp point?

Why is the restricted boltzmann machine both unsupervised and generative?

Comparing approaches of MLE estimates of a Weibull distribution

What problem does Residual Nets solve that batch normalization does not solve?

Energy function of RBM

Factors with only two variables in factor analysis

Approximate mean from .25, .5, .75 percentiles

Exercises in EM with censored data

How many AdaBoost iterations?