Related Experiment Video
Updated: May 25, 2025

An R-Based Landscape Validation of a Competing Risk Model
Published on: September 16, 2022
Direction of impact for explainable risk assessment modeling
Emanuele Borgonovo1, Manel Baucells2, Antonio De Rosa1
1Bocconi University, Milan, Italy.
This study evaluates graphical indicators for model interpretability, finding that only PD functions consistently align with model properties. Analysts should also consider extrapolation risk when choosing visualization tools.
Area of Science:
- Quantitative modeling
- Risk analysis
- Data visualization
Background:
- Graphical indicators aid in visualizing input effects in complex models for decision-makers and risk analysts.
- Limited understanding exists regarding the adequacy and consistency of various marginal effect indicators.
Purpose of the Study:
- To investigate popular marginal effect indicators for consistency with quantitative model properties.
- To examine indicator consistency concerning monotonicity, Lipschitz, and concavity properties.
Main Methods:
- Evaluation of popular marginal effect indicators.
- Assessment of consistency with model properties like monotonicity, Lipschitz, and concavity.
- Consideration of model extrapolation risk.
Main Results:
- Surprisingly, only Probability of Default (PD) functions demonstrated consistency with all examined model properties.
- Individual Conditional Expectations (ICE) plots are recommended when extrapolation risk is manageable.
Conclusions:
- PD functions offer reliable insights into model behavior, satisfying key consistency criteria.
- The choice of visualization indicators must balance model consistency with the risk of extrapolation.
More Related Videos
07:31Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
06:09Low-intensity Blast Wave Model for Preclinical Assessment of Closed-head Mild Traumatic Brain Injury in Rodents
Published on: November 6, 2020
Related Concept Videos
Impact
When particles with different initial velocities collide, they induce deformation by applying equal and opposite impulses. At the point of maximum deformation, the particles move together with...
Types of Biopharmaceutical Studies: Controlled and Non-Controlled Approaches
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast,...
Types of Impact
The coefficient of restitution is a metric for understanding the dynamics of impacts. It quantifies the ratio of relative velocity...
Impact Loading
In cases of elastic deformation,...
Strategies for Assessing and Addressing Confounding
Confounding can be addressed at both the design phase of a study and through analytical methods after data...
Relative Risk