Tool paper (HSB20 CMSB20)

Accepted in HSB20 but not published, Rejected in CMSB20

Things that have been done:

Models, Properties, Rational functions, Data

• Same as in HSB19

Methods

• Space Refinement -

  • Alg. optimisations
  • Code optimisations
  • progress bar

• Qunatitative space sampling NEW

  • so far only implementation
  • sum of L1 distances from dissatisfying constraints in each point (grid sampling)

• Regression NEW

  • single point optimisation using scipy, L2 distance

• Metropolis Hastings NEW

  • Walk in parameter space
  • Bayesian inference to estimate likelihood of a new parameterization
  • burn-in
  • iteration count / timeout
  • gridsize to assign accepted points
  • visualisation
    • 2D
    • line-scatter plot for multidim

Tool NEW

• Gui
• config

Analysis

• MH and Optimisation results added for the models from HSB19

To be done

• Tool part of this
• review here