How to specify Test and Validation data for a SeriesNetwork

1 次查看（过去 30 天）
Hamid Moazed
Hamid Moazed 2018-2-27
编辑： Hamid Moazed 2018-2-27
An answer I found to another question states there is an "older" type of neural network that is a struct, and a "newer" one that is an object. That seems to be the case because much of the documentation I have found during my searches do not seem to apply to what I'm doing. I'm using the documented way of specifying layers and then calling net = createNetwork(X,Y,layers,options) which in my case creates a SeriesNetwork. But I can't figure out how to specify the split for the data between training, validation, and test. I found this in the documentation but it doesn't seem to apply: Divide Data for Optimal Neural Network Training I don't have a network until I call createNetwork() so I can't possibly specify the split using the above method. I've also looked here: Options for training neural network and the only option I found, 'ValidationData', only applies to images, not series data. Any help will be greatly appreciated.

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by