Exploring sensitivity analysis concepts
The variability, or uncertainty, associated with a parameter propagates throughout the model, making it a strong contribution to the variability of the model’s outputs. The model results can be highly correlated with an input parameter so that small changes in the input cause significant changes in the output. A widely used methodology in the field of data analytics is sensitivity analysis. It studies the correlation between the uncertainty of the output of a mathematical model and the various sources of randomness present in the input: we speak of uncertainty analysis when we focus on the quantitative aspect of the problem. There are many objectives of this type of study; here are some examples:
- Understand the complex relationships that exist between the input and output variables
- Identify the most influential risk factors (factor prioritization)
- Check the robustness of the model output to even minor variations of the...
