Skip to contents

Random Walk Prior

Source: R/bage_prior-constructors.R
RW.Rd

Use a random walk as a model for a main effect, or use multiple random walks as a model for an interaction. Typically used with terms that involve age or time.

Usage

RW(s = 1, sd = 1, along = NULL, con = c("none", "by"))

Arguments

s

Scale for the prior for the innovations. Default is 1.

sd

Standard deviation of initial value. Default is 1. Can be 0.

along

Name of the variable to be used as the 'along' variable. Only used with interactions.

con

Constraints on parameters. Current choices are "none" and "by". Default is "none". See below for details.

Value

An object of class "bage_prior_rwrandom" or "bage_prior_rwzero".

Details

If RW2() is used with an interaction, a separate random walk is constructed within each combination of the 'by' variables.

Argument s controls the size of innovations. Smaller values for s tend to produce smoother series.

Argument sd controls variance in initial values. Setting sd to 0 fixes initial values at 0.

Mathematical details

When RW() is used with a main effect,

$$\beta_1 \sim \text{N}(0, \mathtt{sd}^2)$$ $$\beta_j \sim \text{N}(\beta_{j-1}, \tau^2), \quad j > 1$$

and when it is used with an interaction,

$$\beta_{u,1} \sim \text{N}(0, \mathtt{sd}^2)$$ $$\beta_{u,v} \sim \text{N}(\beta_{u,v-1}, \tau^2), \quad v > 1$$

where

  • \(\pmb{\beta}\) is the main effect or interaction;

  • \(j\) denotes position within the main effect;

  • \(v\) denotes position within the 'along' variable of the interaction; and

  • \(u\) denotes position within the 'by' variable(s) of the interaction.

Parameter \(\tau\) has a half-normal prior $$\tau \sim \text{N}^+(0, \mathtt{s}^2),$$ where s is provided by the user.

Constraints

With some combinations of terms and priors, the values of the intercept, main effects, and interactions are are only weakly identified. For instance, it may be possible to increase the value of the intercept and reduce the value of the remaining terms in the model with no effect on predicted rates and only a tiny effect on prior probabilities. This weak identifiability is typically harmless. However, in some applications, such as forecasting, or when trying to obtain interpretable values for main effects and interactions, it can be helpful to increase identifiability through the use of constraints.

Current options for constraints are:

  • "none" No constraints. The default.

  • "by" Only used in interaction terms that include 'along' and 'by' dimensions. Within each value of the 'along' dimension, terms across each 'by' dimension are constrained to sum to 0.

See also

  • RW_Seas() Random walk with seasonal effect

  • RW2() Second-order random walk

  • AR() Autoregressive with order k

  • AR1() Autoregressive with order 1

  • Sp() Smoothing via splines

  • SVD() Smoothing over age using singular value decomposition

  • priors Overview of priors implemented in bage

  • set_prior() Specify prior for intercept, main effect, or interaction

Examples

RW()
#>   RW() 
#>          s: 1
#>         sd: 1
#>      along: NULL
#>        con: none
RW(s = 0.5)
#>   RW(s=0.5) 
#>          s: 0.5
#>         sd: 1
#>      along: NULL
#>        con: none
RW(sd = 0)
#>   RW(sd=0) 
#>          s: 1
#>         sd: 0
#>      along: NULL
#>        con: none
RW(along = "cohort")
#>   RW(along="cohort") 
#>          s: 1
#>         sd: 1
#>      along: cohort
#>        con: none