/MAntRA

Primary LanguageJupyter Notebook

MAntRA

This repository contains code for discovering equations, generating random force, and evaluating reliability of a system. Here are some important notes to keep in mind:

  1. Equation discovery codes for each example are given in the folder with their name. "Drift 1" refers to the code for the first floor, and so on.
  2. To run the reliability code, one needs the prediction data generated by the actual/final prediction file.
  3. A forcing pickle file is used to generate random force.
  4. In the VB algorithm RVM is used for initialization. You must save the Sparsebayes package (to run RVM) in the folder where the codes are being run.
  5. The Sparsebayes package is given in the repository.
  6. The necessary code to execute Gibbs sampling for initialization can be found in the Utils directory.
  7. The order for running the codes is as follows:
    1. Drift & Diffusion files for equation discovery
    2. Forcing pickle file to generate random force
    3. Initial & Final prediction file to generate a larger prediction window for reliability
    4. Reliability file to find the probability of failure.