MATLAB Answers

Help needed to build a binary classifier using neural network to classify electrical faults

2 views (last 30 days)
Neel Soni
Neel Soni on 18 Mar 2020
Answered: Bhargavi Maganuru on 24 Mar 2020
Hello,
I am trying to build a binary classifier to detect High Impedance Faults from simulated data using neural net designer app. I am stuck with how to input the data table in the classifier network. My data consists of Voltage and Current values of all three phases and a Fault Status variable.
These are the steps I am taking:
  1. In the Deep Network Designer app, I added the desired layers and exported the model into workspace.
  2. In the transfer learning app, when I trying to select the .mat file created in the step 1, it shows me an error as below:
Basically, I am stuck with how to import data (Datastore) into the transfer learning app.
Thanks.

  0 Comments

Sign in to comment.

Answers (1)

Bhargavi Maganuru
Bhargavi Maganuru on 24 Mar 2020
Hi,
To import model from mat file, select import mat file model in drop down menu of Pre-Trained model and then press continue in the pop up that comes up. Now it will ask for .mat file, select the corresponding .mat file.
If images stored in folder according to their category, then import only the main folder - using import data(Datastore)
Follow this link for the example.
Hope this helps!

  0 Comments

Sign in to comment.

Sign in to answer this question.