disaggregate_covariate

Disaggregate regional means to fine grid, preserving aggregate means

Supports modelling real-time case data to facilitate the real-time
surveillance of infectious diseases and other point phenomena. The package provides automated computational grid generation over
an area of interest with methods to map covariates between geographies, model fitting including spatially aggregated case counts,
and predictions and visualisation. Both Bayesian and maximum likelihood methods are provided. Log-Gaussian Cox Processes are described by
Diggle et al. (2013) <doi:10.1214/13-STS441> and we provide both the low-rank approximation for Gaussian processes
described by Solin and Särkkä (2020) <doi:10.1007/s11222-019-09886-w> and Riutort-Mayol et al (2023) <doi:10.1007/s11222-022-10167-2> and the
nearest neighbour Gaussian process described by Datta et al (2016) <doi:10.1080/01621459.2015.1044091>.

Sam Watson

rts2

Real-Time Disease Surveillance

disaggregate_covariate function

<dl><dt>x_r</dt>
<dd>length m vector of regional means (or pop-weighted means)</dd>
<dt>W</dt>
<dd>m x n aggregation matrix (1 if cell in region, 0 otherwise)</dd>
<dt>adj</dt>
<dd>n x n adjacency matrix</dd>
<dt>pop</dt>
<dd>optional length n population vector; if provided, preserves
population-weighted means rather than simple means</dd>
<dt>lambda_smooth</dt>
<dd>weight on spatial smoothness</dd>
<dt>lambda_ridge</dt>
<dd>small ridge for numerical stability (needed because L is singular)</dd></dl>

Arguments

Disaggregate regional means to fine grid, preserving aggregate means — disaggregate_covariate

<dl>

<dt>x_r</dt>
<dd>length m vector of regional means (or pop-weighted means)</dd>


<dt>W</dt>
<dd>m x n aggregation matrix (1 if cell in region, 0 otherwise)</dd>


<dt>adj</dt>
<dd>n x n adjacency matrix</dd>


<dt>pop</dt>
<dd>optional length n population vector; if provided, preserves
population-weighted means rather than simple means</dd>


<dt>lambda_smooth</dt>
<dd>weight on spatial smoothness</dd>


<dt>lambda_ridge</dt>
<dd>small ridge for numerical stability (needed because L is singular)</dd>

</dl>

disaggregate_covariate: Disaggregate regional means to fine grid, preserving aggregate means

Description

Usage

Value

Arguments