Sensitivity analysis beyond linearity
From MaRDI portal
Publication:2302770
DOI10.1016/J.IJAR.2019.06.007zbMATH Open1471.62392arXiv1901.02062OpenAlexW2962880903WikidataQ127575555 ScholiaQ127575555MaRDI QIDQ2302770FDOQ2302770
Authors: Manuele Leonelli 
Publication date: 26 February 2020
Published in: International Journal of Approximate Reasoning (Search for Journal in Brave)
Abstract: A wide array of graphical models can be parametrised to have atomic probabilities represented by monomial functions. Such monomial structure has proven very useful when studying robustness under the assumption of a multilinear model where all monomial have either zero or one exponents. Robustness in probabilistic graphical models is usually investigated by varying some of the input probabilities and observing the effects of these on output probabilities of interest. Here the assumption of multilinearity is relaxed and a general approach for sensitivity analysis in non-multilinear models is presented. It is shown that in non-multilinear models sensitivity functions have a polynomial form, conversely to multilinear models where these are simply linear. The form of various divergences and distances under different covariation schemes is also formally derived. Proportional covariation is proven to be optimal in non-multilinear models under some specific choices of varied parameters. The methodology is illustrated throughout by an educational application.
Full work available at URL: https://arxiv.org/abs/1901.02062
Recommendations
- Sensitivity analysis in multilinear probabilistic models
- Sensitivity analysis of probabilistic networks
- Model-preserving sensitivity analysis for families of Gaussian distributions
- Co-variation for sensitivity analysis in Bayesian networks: properties, consequences and alternatives
- Local Sensitivity of Bayesian Networks to Multiple Simultaneous Parameter Shifts
Cites Work
- Conditional independence and chain event graphs
- Title not available (Why is that?)
- On Information and Sufficiency
- Probabilistic graphical models.
- Title not available (Why is that?)
- Bayesian decision analysis. Principles and practice.
- A Differential Approach for Staged Trees
- A differential approach to inference in Bayesian networks
- Bayesian MAP model selection of chain event graphs
- A distance measure for bounding probabilistic belief change
- Title not available (Why is that?)
- The effect of block parameter perturbations in Gaussian Bayesian networks: sensitivity and robustness
- Bayesian hidden Markov model for DNA sequence segmentation: a prior sensitivity analysis
- An extension of the differential approach for Bayesian network inference to dynamic Bayesian networks
- Properties of sensitivity analysis of Bayesian belief networks
- Sensitivity analysis in multilinear probabilistic models
- A simplified matrix formulation for sensitivity analysis of hidden Markov models
- Efficient sensitivity analysis in hidden Markov models
- Co-variation for sensitivity analysis in Bayesian networks: properties, consequences and alternatives
- A symbolic algebra for the computation of expected utilities in multiplicative influence diagrams
Cited In (7)
- A geometric characterization of sensitivity analysis in monomial models
- New sensitivity analysis subordinated to a contrast
- Editorial. Special issue on robustness in probabilistic graphical models
- Title not available (Why is that?)
- Title not available (Why is that?)
- Automatically finding the right probabilities in Bayesian networks
- The YODO algorithm: an efficient computational framework for sensitivity analysis in Bayesian networks
Uses Software
This page was built for publication: Sensitivity analysis beyond linearity
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2302770)
