How to assign labels to a dataset using 'signalDatastore' object?

Question

BIPIN SAMUEL 2024-6-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2131461-how-to-assign-labels-to-a-dataset-using-signaldatastore-object

评论： BIPIN SAMUEL 2024-6-25

I have a dataset in a folder for classification purpose using crossentropy as loss function in deep learning. I need to add labels to my dataset. How can I do this? Whether signalDatastore object helps to assign labels like imageDatastore object?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

dpb 2024-6-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2131461-how-to-assign-labels-to-a-dataset-using-signaldatastore-object#answer_1476406

Not at all sure about what you mean by "object helps to assign labels like imageDatastore object".

signalDatastore <example> illustrates picking two variables out of a set of files; it does require you know what the desired variable names from the files are, but that seems like something would have to know, anyway.

But, there's also an example of programmatically return variable names from a .mat or text file under the SignalVariableNames parameter for files.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

BIPIN SAMUEL 2024-6-25

Thank You @dpb for the response.

ds = imageDatastore(filefolder,...

'IncludeSubfolders',true,...

'LabelSource','foldernames');

Here the labels are assinged using the 'foldernames'. So, we are not needed to provide the target data (labels) separately. Is it possible to use "signalDatastore" or "arrayDatastore" like this? So, that we can use the network for classification using crossentropy as loss function. Moreover, I have one more query. Whether it is possible to use the function trainNetwork() to train a network with two inputs i.e, one with sequential data and other with its features. Like an example shown here. But here trainnet() is used for training. Since I am using Matlab 2023a version this function is not available.

请先登录，再进行评论。