Can I use Deep Network Designer to classify data in a table for fault Classification.

Fault Diagnosis of Centrifugal Pumps Using Residual Analysis - MATLAB & Simulink (mathworks.com)

Using the code from here I want to take the FeaturesTable and insert it into a Deep Network Designer to combine SVM and other network classification methods. It seems that the Deep Network Designer only wants images as inputs. Is there a way to input data tables in for the fault classification?

This is what pops up when I try to import data:

This is a quick screenshot of what the data table looks like.

Answer 1

[Shoaib iqbal]

Hi Lawrence,

I understand that you are trying to input a table to Deep Learning model using Deep Learning Designer.

To input the table, use the layer 'featureInputLayer'. Then you need to create a datastore from your input. In the workspace, you can write the command

ds = arrayDatastore(tbl); %tbl is your table

After creating, you can see the same datastore in "Import DataStore" tab.

For more information on 'featureInputLayer' and 'Train network with Numeric Features', refer to the docs

https://in.mathworks.com/help/deeplearning/ug/train-network-on-data-set-of-numeric-features.html

https://in.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.featureinputlayer.html

Hope this solves your query.