Unsure

A simple app to efficiently account for multiple independent sources of uncertainty in complex predictive models

---

Unsure approximates the user's model with a Cobb-Douglas function to rapidly estimate results with any combination of input values. It requires that the user already has a working model and has run a standard sensitivity analysis.

Inputs

inputDistributions.csv is a table of the potential values of uncertain inputs, and their associated probabilities

• id - ID of the input variable
• name - name of the input variable
• value - value of the input variable
• probability - probability that input variable has this value

inputElasticities.csv is a table of the results of a sensitivity analysis of the user's model

• id - ID of the input variable
• name - name of the input variable
• base_value - default value of input variable
• elasticity - elasticity of model output with respect to input variable

Run

1. Run server.R or ui.R in R
2. Navigate to default port in a browser
3. Upload inputDistributions.csv and inputElasticities.csv
4. View Cumulative Distribution Function (CDF) and table (Stats) of potential results
5. Select input variables and values under Explorer to highlight filtered results
6. Download CSV of results under Output