Stratify curve simulation by severity

[jrgant]

Need to stratify the curve simulations to produce two sets of hypothetical influenza hospitalization curves: one for High/Moderate and on for Low severity seasons.

[jrgant]

Stratification has been implemented.

However, simulated curves from Low severity seasons somehow have a higher mean hospitalization count than High/Moderate.

Possible issues:

• Bug that results in function not correctly sampling the specified severity type
• Issue with the trend filter predictions produced by each historical season (lambda value or other issue)
• Only 4 low-severity seasons available... too much variance?

[jrgant]

We'll be moving this to a sensitivity analysis. The comment above (#2 (comment)) has been addressed.