This repository is based on our paper titled "Machine Learning-Assisted Equivalent Circuit Identification for Dielectric Spectroscopy of Polymers." The paper discusses a novel convolutional neural network (CNN) model to predict the electrical equivalent circuit (EEC) topology from broadband dielectric spectroscopy data, enhancing the characterization of polymer membranes' and achieving SOTA resutls.
To install the required packages, please run the following command:
pip install -r requirements.txt
You can easliy run the model in main.py
and define global parameters.
The file contains the modules required for generating the data model.generate_data()
and pre-processing the data model.preprocess_data(data=data)
.
For training the model, the user can call model.Train(Training_data)
, and
for testing the model you can use model.Predict(Test_data)
.
Addtionally model.Predict(Test_data)
will save the results in the folder "predictions".