Skip to main content
Skander H.'s user avatar
Skander H.'s user avatar
Skander H.'s user avatar
Skander H.
  • Member for 8 years, 8 months
  • Last seen more than a week ago
  • Seattle, WA, USA
40 votes
7 answers
9k views

Why doesn't regularization solve Deep Neural Nets hunger for data?

35 votes
5 answers
5k views

How to handle a "self defeating" prediction model?

19 votes
1 answer
4k views

If an auto-regressive time series model is non-linear, does it still require stationarity?

19 votes
4 answers
829 views

On George Box, Galit Shmueli and the scientific method?

15 votes
4 answers
24k views

Statsmodels says ARIMA is not appropriate because series is not stationary, how is it testing that?

13 votes
1 answer
9k views

Why is Maximum Likelihood used for ARIMA instead of least squares?

12 votes
2 answers
3k views

What ever happened to Fuzzy Logic?

11 votes
2 answers
17k views

How can a network with only ReLU nodes output negative values?

11 votes
3 answers
245 views

Is there a way to recover a temporal dependence structure in a time series from a regression against time?

10 votes
1 answer
6k views

When is the AIC a good model selection criterion for forecasting and when is it not?

10 votes
4 answers
6k views

Does the universal approximation theorem for neural networks hold for any activation function?

9 votes
1 answer
2k views

How can the AIC or BIC be used instead of the train/test split?

9 votes
2 answers
3k views

If the AIC and the BIC are asymptotically equivalent to cross validation, is it possible to dispense with a test set when using them?

9 votes
1 answer
2k views

What can Deep Neural Networks do that Support Vector Machines can't?

9 votes
1 answer
2k views

What is the difference between model selection and hyperparameter tuning?

9 votes
2 answers
2k views

Why would somebody use a hash function for creating a test/train split instead of random seed?

8 votes
0 answers
4k views

Which loss function to use when training LSTM for time series?

8 votes
1 answer
2k views

How to you measure the accuracy of a model that gives quantile forecasts or distributions of forecasts?

8 votes
1 answer
5k views

Can one force an ARIMA forecast to be positive?

6 votes
1 answer
1k views

Is there a difference between 1D Mean Shift and KDE for clustering 1 d data?

6 votes
1 answer
2k views

Why are exponential smoothing models not considered auto-regressive?

6 votes
1 answer
7k views

How does logistic regression "elegantly" handle unbalanced classes?

5 votes
1 answer
47 views

Generating embeddings for languages without a written representation?

5 votes
0 answers
355 views

Is there a theoretical reason why simple models perform better than complex models on time series forecasting tasks?

5 votes
2 answers
966 views

The Universal Approximation Theorem vs. The No Free Lunch Theorem: What's the caveat?

5 votes
1 answer
2k views

How does Ridge Regression penalize for complexity if the coefficients are never allowed to go to zero?

5 votes
1 answer
830 views

How is an RNN (or any neural network) a parametric model?

5 votes
1 answer
462 views

Why is removing instationarities a good thing when trying to forecast a time series?

5 votes
1 answer
265 views

How versatile is logistic regression?

5 votes
2 answers
899 views

auto.arima returns a non-seasonal model even though I am forcing seasonality