The models created with functions mod_pois(), mod_binom(),
and mod_norm() always include an intercept,
and typically include main effects
and interactions formed from variables in input data.
Most models, for instance
include an age effect, and many include
an interaction between age and sex/gender,
or age and time.
The intercept, main effects, and interactions all have prior models that capture the expected behavior of the term. Current choices of prior models are summarised here.
Details
| Prior | Description | Uses | Forecast |
N() | Elements drawn from normal distribution | Term with no natural order | Yes |
NFix() | As for N(), but standard deviation fixed | Term with few elements | Yes |
RW() | Random walk | Smoothing | Yes |
RW2() | Second-order random walk | Like RW(), but smoother | Yes |
RW_Seas() | Random walk, with seasonal effect | Terms involving time | Yes |
RW2_Seas() | Second-order random walk, with seasonal effect | Term involving time | Yes |
AR() | Auto-regressive prior of order k | Mean reversion | Yes |
AR1() | Auto-regressive prior of order 1 Special case of AR() | Mean reversion | Yes |
Known() | Values treated as known | Simulations, prior knowledge | No |
Lin() | Linear trend, with independent normal | Parsimonious model for time | Yes |
Lin_AR() | Linear trend, with autoregressive errors | Term involving time | Yes |
Lin_AR1() | Linear trend, with AR1 errors | Terms involving time | Yes |
Sp() | P-Spline (penalised spline) | Smoothing, eg over age | No |
SVD() | Age or age-sex profile based on SVD of database | Age or age-sex | No |
SVD_AR() | SVD(), but coefficients follow AR() | Age or age-sex and time | Yes |
SVD_AR1() | SVD(), but coefficients follow AR1() | Age or age-sex and time | Yes |
SVD_RW() | SVD(), but coefficients follow RW() | Age or age-sex and time | Yes |
SVD_RW2() | SVD(), but coefficients follow RW2() | Age or age-sex and time | Yes |