I have used Multi-Layer Perceptron Neural Networks to do a binary classification. Now I want to perform a sensitivity analysis. I am planning to follow a similar approach to this paper.
In case the paper is not assessable, below is the part of the paper where they have implemented the sensitivity analysis (GESM is a short form that they have used for gully erosion susceptibility maps).
I am doing the analysis in python and have saved the model in my system. I am able to understand what they are doing in their work, but not able to understand how it is being implemented in python. I came upon one library for the same called Sensitivity Analysis Library (SALib). But I am not able to understand how to give the saved model to this library and do the analysis.
Is there any standard method to do the sensitivity analysis in python?
My dataset has 28 features which are a combination of Continuous, Discrete & Categorical Variables.