postPhi

In our model the data are drawn from LogN(mu_i + log(c_ij), tau_i). The prior for tau_i
is given as Gamma(thetaiphi, phi). This function uses a MH step to draw a new sample of phi.
Proposal distribution is Gamma(currentPhirhoPhi, rhoPhi).
Note that we parameterize with RATE, not SCALE.

Implements a Bayesian hierarchical model designed to identify skips in mobile menstrual cycle self-tracking on mobile apps. Future developments will allow for the inclusion of covariates affecting cycle mean and regularity, as well as extra information regarding tracking non-adherence. Main methods to be outlined in a forthcoming paper, with alternative models from Li et al. (2022) <doi:10.1093/jamia/ocab182>.

Luke Duttweiler

skipTrack

A Bayesian Hierarchical Model that Controls for Non-Adherence in
Mobile Menstrual Cycle Tracking

postPhi function

<dl><dt>taui</dt>
<dd>Numeric vector, individuals precisions.</dd>
<dt>thetai</dt>
<dd>Numeric vector. individuals precisions means (estimate)</dd>
<dt>currentPhi</dt>
<dd>Previous draw of phi</dd>
<dt>rhoPhi</dt>
<dd>Proposal rate for gamma distribution that draws proposal for phi, default is 1000.</dd></dl>

Arguments

Metropolis-Hastings step to draw a new value for phi. — postPhi

<dl>

<dt>taui</dt>
<dd>Numeric vector, individuals precisions.</dd>


<dt>thetai</dt>
<dd>Numeric vector. individuals precisions means (estimate)</dd>


<dt>currentPhi</dt>
<dd>Previous draw of phi</dd>


<dt>rhoPhi</dt>
<dd>Proposal rate for gamma distribution that draws proposal for phi, default is 1000.</dd>

</dl>

postPhi: Metropolis-Hastings step to draw a new value for phi.

Description

Usage

Value

Arguments