search

jrgant / FluHospPrediction

R package repository to reproduce the flu hospitalization prediction project results
1 stars 0 forks source link

Error Distribution #4

Closed jrgant closed 5 years ago

jrgant commented 5 years ago

My instinct is that we should still figure out an appropriate transformation for the error distribution (currently Gaussian with mean 0). Otherwise, any negative hospitalization count produced by f(i) + ε, ε ~ N(μ, σ^2) gets transformed to 0, and we lose the random error.

I suspect the goal should be this: 0.5 * [|f(i)| + f(i)] + ε, ε ~ dist(x).

On the other hand, we know counts during the early season will be near zero.

jrgant commented 5 years ago

I'm thinking this is not actually a problem anymore. Curves look pretty good within the range of weeks we're interested in.