Sensitivity Analysis quantifies the impact of variable changes on a specific outcome within a model. Employed across various disciplines, it aids in risk assessment, model validation, and decision-making, offering metrics to represent sensitivity.

## Definition

Quantitative technique to assess how different values of an independent variable impact a particular dependent variable under a given set of assumptions.

## Types of Variables

**Input Variables:**Factors being manipulated in the analysis.**Output Variables:**Resultant variables that change based on input variations.

## Types of Sensitivity Analysis

**Local Sensitivity Analysis:**Examines change over a small range of parameter values.**Global Sensitivity Analysis:**Covers a wide range of parameter values.**Deterministic:**Uses specific set values for inputs.**Probabilistic:**Incorporates randomness in inputs and/or outputs.

## Methods

**One-at-a-Time (OAT):**Changes one variable while keeping others constant.**Monte Carlo Simulation:**Uses random sampling to obtain numerical results.**Factorial Analysis:**Investigates the effects of multiple variables at once.

## Applications

**Risk Assessment:**Used to estimate uncertainties in outcomes.**Model Validation:**Helps in refining models by comparing results to real-world data.**Decision Support:**Assists in choosing between different strategies or scenarios.

## Metrics and Indicators

**Elasticity:**Measure of sensitivity, often expressed as a percentage change.**Tornado Diagrams:**Graphical representation ranking variables by their impact.**Sobol Indices:**Quantify the contribution of each input to the output variance.

## Limitations

**Computational Complexity:**Especially relevant for high-dimensional models.**Assumptions:**Results as good as the assumptions they are based on.

## Interdisciplinary Usage

**Finance:**Option pricing, portfolio optimization.**Engineering:**System reliability, material selection.**Medicine:**Epidemiological models, treatment effectiveness.

## Ethical Considerations

**Transparency:**Clear methodology essential for validity.**Misuse:**Risk of cherry-picking data or manipulating variables for desired outcomes.