User jpmuc

115 votes

Accepted

tanh activation function vs sigmoid activation function

answered Jun 8, 2014 at 6:47

70 votes

Gradient Boosting Tree vs Random Forest

answered Feb 13, 2016 at 14:46

34 votes

Accepted

What is one class SVM and how does it work?

answered May 18, 2014 at 20:16

32 votes

Accepted

What does kernel size mean?

answered Aug 7, 2017 at 20:43

20 votes

Accepted

Is KNN a discriminative learning algorithm?

answered Jul 7, 2014 at 16:50

19 votes

Accepted

Where is density estimation useful?

answered Feb 3, 2014 at 20:01

16 votes

Machine learning algorithms to handle missing data

answered Jun 16, 2014 at 11:21

15 votes

Accepted

Why not always use ensemble learning?

answered Jul 29, 2014 at 13:52

14 votes

Why is optimizing a mixture of Gaussian directly computationally hard?

answered Apr 21, 2014 at 16:16

14 votes

Softmax layer in a neural network

answered Dec 12, 2013 at 22:29

14 votes

Accepted

How to calculate optimal zero padding for convolutional neural networks?

answered Aug 13, 2017 at 12:05

14 votes

Accepted

Why is gradient descent required?

answered May 16, 2016 at 0:40

14 votes

Accepted

Curse of dimensionality: kNN classifier

answered Jun 28, 2015 at 21:42

13 votes

Accepted

Having a conjugate prior: Deep property or mathematical accident?

answered Aug 29, 2013 at 8:03

12 votes

Does k-NN with k=1 always implies overfitting?

answered Jul 14, 2014 at 17:34

11 votes

What are good initial weights in a neural network?

answered Jan 13, 2013 at 20:45

11 votes

Accepted

Why must one trade off between bias and variance?

answered Mar 26, 2013 at 16:42

10 votes

Accepted

Mathematical background for neural networks

answered Jul 15, 2013 at 13:06

10 votes

Machine Learning to Predict Class Probabilities

answered Nov 16, 2013 at 12:02

10 votes

Accepted

Iterative process for removing extreme samples

answered Jul 25, 2014 at 14:24

10 votes

Accepted

Non-orthogonal technique analogous to PCA

answered May 9, 2014 at 12:59

10 votes

Regression vs. classification and generative vs. discriminative

answered Jun 17, 2017 at 17:32

10 votes

Choosing filter size, strides etc. in a CNN?

answered Aug 3, 2017 at 17:49

9 votes

What machine learning techniques can, once trained, generate prediction despite some missing inputs?

answered May 17, 2013 at 14:11

8 votes

Accepted

Visualizing high dimensional data

answered Apr 19, 2013 at 20:56

8 votes

LibSVM cost weights for unbalanced data doesn't work

answered Jan 17, 2013 at 21:17

8 votes

Accepted

Differences between Multi-layer NN, Hopfield, Helmholtz and Boltzmann machines

answered Jan 20, 2014 at 18:02

8 votes

Accepted

stick breaking model of Dirichlet process

answered Apr 15, 2014 at 13:58

7 votes

Naive Bayes: Imbalanced Dataset in Real-time Scenario

answered Jul 4, 2014 at 13:20

7 votes

In Naive Bayes, why bother with Laplace smoothing when we have unknown words in the test set?

answered Jul 22, 2014 at 8:33

Stack Exchange Network

252 Answers

tanh activation function vs sigmoid activation function

Gradient Boosting Tree vs Random Forest

What is one class SVM and how does it work?

What does kernel size mean?

Is KNN a discriminative learning algorithm?

Where is density estimation useful?

Machine learning algorithms to handle missing data

Why not always use ensemble learning?

Why is optimizing a mixture of Gaussian directly computationally hard?

Softmax layer in a neural network

How to calculate optimal zero padding for convolutional neural networks?

Why is gradient descent required?

Curse of dimensionality: kNN classifier

Having a conjugate prior: Deep property or mathematical accident?

Does k-NN with k=1 always implies overfitting?

What are good initial weights in a neural network?

Why must one trade off between bias and variance?

Mathematical background for neural networks

Machine Learning to Predict Class Probabilities

Iterative process for removing extreme samples

Non-orthogonal technique analogous to PCA

Regression vs. classification and generative vs. discriminative

Choosing filter size, strides etc. in a CNN?

What machine learning techniques can, once trained, generate prediction despite some missing inputs?

Visualizing high dimensional data

LibSVM cost weights for unbalanced data doesn't work

Differences between Multi-layer NN, Hopfield, Helmholtz and Boltzmann machines

stick breaking model of Dirichlet process

Naive Bayes: Imbalanced Dataset in Real-time Scenario

In Naive Bayes, why bother with Laplace smoothing when we have unknown words in the test set?