User Haitao Du

79

110

votes

6

answers

40k

views

Is it possible to train a neural network without backpropagation?

Sep 25 '18 at 18:09 Oren Milman 932

16

39

votes

10

answers

8k

views

Why are survival times assumed to be exponentially distributed?

distributions survival assumptions exponential-distribution

Jul 10 at 13:55 klumbard 1,195

24

38

votes

4

answers

5k

views

Is LSTM (Long Short-Term Memory) dead?

machine-learning natural-language lstm sequence-analysis

Jul 4 at 20:14 OverLordGoldDragon 458

9

33

votes

1

answer

2k

views

If I generate a random symmetric matrix, what's the chance it is positive definite?

probability matrix random-generation eigenvalues random-matrix

Dec 4 '18 at 7:01 Ben 65.9k

9

30

votes

1

answer

12k

views

Can degrees of freedom be a non-integer number?

r degrees-of-freedom generalized-additive-model

Dec 27 '18 at 1:32 Glen_b 245k

21

29

votes

3

answers

16k

views

What are the impacts of choosing different loss functions in classification to approximate 0-1 loss

machine-learning classification optimization loss-functions

Apr 20 '18 at 14:53 galoosh33 2,072

37

26

votes

7

answers

15k

views

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

mathematical-statistics optimization covariance-matrix intuition linear-algebra

Oct 22 '19 at 21:33 kjetil b halvorsen♦ 53.3k

20

26

votes

2

answers

9k

views

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

machine-learning continuous-data feature-engineering binning

Sep 27 '17 at 22:07 Alexis 22.3k

6

26

votes

1

answer

2k

views

What are the classical notations in statistics, linear algebra and machine learning? And what are the connections between these notations?

machine-learning probability self-study optimization

Aug 29 '16 at 13:53 Community♦ 1

26

25

votes

2

answers

4k

views

Why there are two different logistic loss formulation / notations?

logistic generalized-linear-model notation loss-functions

Jun 2 '17 at 15:12 Haitao Du 30.2k

23

votes

1

answer

6k

views

Is there any intuitive explanation of why logistic regression will not work for perfect separation case? And why adding regularization will fix it?

logistic generalized-linear-model optimization intuition separation

Apr 27 at 17:14 ady 103

16

21

votes

1

answer

6k

views

Why using Newton's method for logistic regression optimization is called iterative re-weighted least squares?

logistic generalized-linear-model optimization irls fisher-scoring

May 8 '18 at 15:43 jld 15.9k

9

20

votes

2

answers

4k

views

Are line search methods used in deep learning? Why not?

machine-learning neural-networks optimization deep-learning

Aug 7 '18 at 17:38 Derrell D'Souza 1

19

votes

3

answers

24k

views

What exactly is the difference between a parametric and non-parametric model?

machine-learning neural-networks terminology nonparametric

Nov 11 at 12:52 Lavanya Sagunthala 21

7

19

votes

3

answers

25k

views

What is a null model in regression and how does it relate to the null hypothesis?

regression hypothesis-testing classification terminology model

Nov 9 at 14:16 Lerner Zhang 3,696

16

19

votes

2

answers

3k

views

What is happening here, when I use squared loss in logistic regression setting?

r machine-learning logistic optimization loss-functions

Nov 1 at 5:28 Haitao Du 30.2k

13

19

votes

2

answers

9k

views

Is there i.i.d. assumption on logistic regression?

regression logistic assumptions iid

May 4 at 9:13 Haitao Du 30.2k

2

19

votes

3

answers

4k

views

Is a decision stump a linear model?

machine-learning cart linear boosting

Mar 19 '17 at 1:56 Community♦ 1

9

18

votes

2

answers

12k

views

Matrix notation for logistic regression

regression logistic linear-model notation

Dec 8 '19 at 12:30 Community♦ 1

9

18

votes

4

answers

4k

views

Does “curse of dimensionality” really exist in real data?

clustering dimensionality-reduction high-dimensional

Jun 23 '16 at 1:30 Community♦ 1

8

17

votes

3

answers

5k

views

Why the default matrix norm is spectral norm and not Frobenius norm?

matrix linear-algebra

Feb 20 '19 at 2:47 dcneuro 5

11

15

votes

3

answers

4k

views

Is PCA optimization convex?

machine-learning pca optimization convex

Jan 19 at 22:45 whuber♦ 258k

5

15

votes

2

answers

3k

views

Why linear regression has assumption on residual but generalized linear model has assumptions on response?

regression generalized-linear-model assumptions linear

Feb 24 '18 at 2:27 Cliff AB 17k

7

15

votes

2

answers

2k

views

Why Expectation Maximization is important for mixture models?

machine-learning optimization expectation-maximization gaussian-mixture-distribution

Dec 27 '17 at 2:50 dontloo 12k

6

14

votes

1

answer

14k

views

Can support vector machine be used in large data?

machine-learning svm large-data

Jul 26 at 21:22 ikamen 103

17

13

votes

5

answers

16k

views

When to use mixed effect model?

regression mixed-model random-effects-model linear

Jun 17 '18 at 17:14 user211885 1

6

13

votes

2

answers

3k

views

Is R's glm function useless in a big data / machine learning setting? [duplicate]

r logistic generalized-linear-model glmnet separation

Oct 12 '16 at 12:43 Community♦ 1

11

votes

1

answer

553

views

Why does R have a different definition on convolution?

r python convolution history

May 7 at 20:53 Community♦ 1

4

10

votes

3

answers

1k

views

Difference between MLE and Baum Welch on HMM fitting

hidden-markov-model expectation-maximization

Oct 5 '18 at 13:24 Lerner Zhang 3,696

1

10

votes

1

answer

1k

views

Why I am getting different predictions for manual polynomial expansion and using the R `poly` function?

r regression polynomial

May 2 '17 at 4:51 Community♦ 1

Stack Exchange Network

current community

your communities

more stack exchange communities

Next badge

194 Questions

Is it possible to train a neural network without backpropagation?

Why are survival times assumed to be exponentially distributed?

Is LSTM (Long Short-Term Memory) dead?

If I generate a random symmetric matrix, what's the chance it is positive definite?

Can degrees of freedom be a non-integer number?

What are the impacts of choosing different loss functions in classification to approximate 0-1 loss

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

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

What are the classical notations in statistics, linear algebra and machine learning? And what are the connections between these notations?

Why there are two different logistic loss formulation / notations?

Is there any intuitive explanation of why logistic regression will not work for perfect separation case? And why adding regularization will fix it?

Why using Newton's method for logistic regression optimization is called iterative re-weighted least squares?

Are line search methods used in deep learning? Why not?

What exactly is the difference between a parametric and non-parametric model?

What is a null model in regression and how does it relate to the null hypothesis?

What is happening here, when I use squared loss in logistic regression setting?

Is there i.i.d. assumption on logistic regression?

Is a decision stump a linear model?

Matrix notation for logistic regression

Does “curse of dimensionality” really exist in real data?

Why the default matrix norm is spectral norm and not Frobenius norm?

Is PCA optimization convex?

Why linear regression has assumption on residual but generalized linear model has assumptions on response?

Why Expectation Maximization is important for mixture models?

Can support vector machine be used in large data?

When to use mixed effect model?

Is R's glm function useless in a big data / machine learning setting? [duplicate]

Why does R have a different definition on convolution?

Difference between MLE and Baum Welch on HMM fitting

Why I am getting different predictions for manual polynomial expansion and using the R `poly` function?