Timeline for How many observations in a neural networks dataset?
Current License: CC BY-SA 3.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 18, 2019 at 1:01 | comment | added | baxx |
@Azrael could you provide an example of a neural network and the number of observations you might expect based on the number of parameters in the neural net? For example, if one was predicting happiness based on hair colour, there could be three levels {black,brown,blond} to the categorical variable hair-colour, seeing how many observations of each are required for a particular neural net would be useful to this answer i think
|
|
| Jun 25, 2015 at 10:17 | comment | added | Azrael | This edit should address your concerns. | |
| Jun 25, 2015 at 10:16 | history | edited | Azrael | CC BY-SA 3.0 |
added more details
|
| Jun 25, 2015 at 8:04 | comment | added | image_doctor | And your first para presents a rule, not a rule of thumb ? | |
| Jun 24, 2015 at 23:43 | comment | added | Azrael | Again, the question asks for a thumb rule. | |
| Jun 24, 2015 at 23:40 | comment | added | image_doctor | I think that demonstrates that your opening paragraph is not true ? | |
| Jun 24, 2015 at 21:38 | comment | added | Azrael | Yes, you are right. For example, you can keep giving a neural network 0 as an input, and 1 as an output multiple times and it will start learning to become a NOT gate. But what the question has demanded is a rule of thumb. The number of parameters that are required can not be correctly predicted without knowing more details. | |
| Jun 24, 2015 at 21:33 | comment | added | image_doctor | Maybe someone else knows better, but I"m not entirely sure this is true. You aren't trying to solve the equations of the neural network in a classical sense. You could quite easily have a complex network that was trained to recognise the difference between only two data points. | |
| Jun 24, 2015 at 21:22 | comment | added | Azrael | No problem. Happy to help. | |
| Jun 24, 2015 at 21:22 | vote | accept | Quantopik | ||
| Jun 24, 2015 at 21:17 | comment | added | Azrael | The data points mean the observations. | |
| Jun 24, 2015 at 21:05 | comment | added | Quantopik | Firstly, thanks for the answer and explanation (+1), @Azrael. So, according to you, shouldn't I care about the number of observation? Could you provide some reference about that? | |
| Jun 24, 2015 at 20:52 | history | answered | Azrael | CC BY-SA 3.0 |