Code supporting the paper: 1D neural network
This online resource shows two archived folders: Matlab and Python, that contain relevant code for the article: A Bayesian finite-element trained machine learning approach for predicting post-burn contraction.
One finds the codes used to generate the large dataset within the Matlab folder. Here, the file Main.m is the main file and from there, one can run the Monte Carlo simulation. There is a README file.
Within the Python folder, one finds the codes used for training the neural networks and creating the online application. The file Data.mat contains the data generated by the Matlab Monte Carlo simulation. The files run_bound.py, run_rsa.py, and run_tse.py train the neural networks, of which the best scoring ones are saved in the folder Training. The DashApp folder contains the code for the creation of the Application.
- 2022-10-28 first online, published, posted
- Dutch Burn Foundation, project 17.105
University of Hasselt, Department of Mathematics and Statistics
- download all files (zip)
276,091,831 bytes unzipped