Timeline for Number of inputs much greater than size of training (Neural Networks)
Current License: CC BY-SA 3.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 8, 2017 at 15:04 | comment | added | Haitao Du | @TommasoGuerrini adding regularization and reducing input dimensions are different things. that is why I said, the model will work well with huge number of inputs (with regularization.) | |
| Mar 8, 2017 at 15:03 | comment | added | Tommaso Guerrini | That's why I thought about reducing dimensionality of the input through an autoencoder. | |
| Mar 8, 2017 at 15:01 | comment | added | Haitao Du | @TommasoGuerrini yes, you could. And you can (and should) use regularization on your model. | |
| Mar 8, 2017 at 15:00 | comment | added | Tommaso Guerrini | Sorry for bothering you again: couldn't I have a pretty high probability of overfitting? | |
| Feb 26, 2017 at 13:24 | vote | accept | Tommaso Guerrini | ||
| Feb 16, 2017 at 16:31 | history | edited | Haitao Du | CC BY-SA 3.0 |
added 20 characters in body
|
| Feb 16, 2017 at 16:29 | vote | accept | Tommaso Guerrini | ||
| Feb 26, 2017 at 13:24 | |||||
| Feb 16, 2017 at 16:14 | history | edited | Haitao Du | CC BY-SA 3.0 |
deleted 61 characters in body
|
| Feb 16, 2017 at 16:09 | history | edited | Haitao Du | CC BY-SA 3.0 |
deleted 61 characters in body
|
| Feb 16, 2017 at 16:07 | comment | added | Haitao Du | @TommasoGuerrini I am trying to say "high number of inputs/features is not an issue whichever the number of data" | |
| Feb 16, 2017 at 16:07 | comment | added | Tommaso Guerrini | By the way (I wrote more here if you are interested: stats.stackexchange.com/questions/260444/…) | |
| Feb 16, 2017 at 16:06 | comment | added | Tommaso Guerrini | Sorry for bothering you, it's not that clear to me one thing: are you saying that the number of inputs ($7k$ in my case) is not necessarily equal to the number of features one extract from those inputs? Or are you plainly saying that a high number of inputs/features is not an issue whichever the number of data? I'm forecasting $2k$ products daily sales. For now I'm considering $340$ products (read output) and a corresponding $\sim 1k$ inputs. Forecasting performance minimizing WeightedMape out performs every other algorithm, so I was doubting the suggestion given to me | |
| Feb 16, 2017 at 16:00 | history | answered | Haitao Du | CC BY-SA 3.0 |