markjrieke / 2024-potus

repo for constructing a 2024 presidential election forecast
MIT License
45 stars 6 forks source link

Investigate prior distributions #27

Open markjrieke opened 2 months ago

markjrieke commented 2 months ago

I set the priors over parameters to restrict the range of plausible values but also to avoid the model from relying on $\beta_n$ too much. Need to investigate a possible loosening of priors, as well as non-normal distributions to account for outliers (i.e., $\eta_n$ may be more t-shaped than normal)