Skip to content

Commit

Permalink
Updates to Appendix A including lengthscale prior sensitivity table
Browse files Browse the repository at this point in the history
  • Loading branch information
@athowes
athowes committed Dec 19, 2023
1 parent 1dd997e commit e7b1cc9
Show file tree
Hide file tree
Showing 12 changed files with 426 additions and 361 deletions.
17 changes: 14 additions & 3 deletions 90-appendixA.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -51,14 +51,25 @@ knitr::include_graphics(path)

## Lengthscale prior sensitivity {#lengthscale-prior}

(ref:lengthscale-prior) Six lengthscale prior distributions were considered. These were: $l \sim \text{Gamma}(1, 1)$; $l \sim \text{IG}(a, b)$; $l \sim \mathcal{N}^{+}(0, \sigma)$; $l \sim \text{Log-normal}(0, 1)$; $p(l) = 1$; and $l \sim \mathcal{N}(2.5, 1)$.
| Description | Prior | Additional details |
|:------------|:------|:--------------------|
| Gamma | $l \sim \text{Gamma}(1, 1)$ ||
| Geometry-informed inverse-gamma | $l \sim \text{IG}(a, b)$ | |
| Geometry-informed normal | $l \sim \mathcal{N}^{+}(0, \sigma)$ | |
| Log-normal | $l \sim \text{Log-normal}(0, 1)$ | |
| Non-informative | $p(l) = 1$ | |
| Oracle normal | $l \sim \mathcal{N}(2.5, 1)$ | |

Table: (\#tab:lengthscale-prior) Six lengthscale prior distributions were considered for use.

(ref:lengthscale-prior) The density for each lengthscale prior distributions as given in Table \@ref(tab:lengthscale-prior).

```{r lengthscale-prior, fig.cap="(ref:lengthscale-prior)"}
path <- paste0("resources/beyond-borders/", resource_version, "/depends/lengthscale-prior.png")
knitr::include_graphics(path)
```

(ref:lengthscale-posterior) Lengthscale posterior distributions obtained using NUTS to fit a centroid kernel model to integrated kernel data, with six different lengthscale prior distributions as described in Figure \@ref(fig:lengthscale-prior).
(ref:lengthscale-posterior) Lengthscale posterior distributions obtained using NUTS to fit a centroid kernel model to integrated kernel data grid on geometry (Panel \@ref(fig:geometries)E), with six different lengthscale prior distributions as given in Table \@ref(tab:lengthscale-prior).

```{r lengthscale-posterior, fig.cap="(ref:lengthscale-posterior)"}
path <- paste0("resources/beyond-borders/", resource_version, "/depends/lengthscale-posterior.png")
Expand All @@ -69,7 +80,7 @@ knitr::include_graphics(path)

### Lengthscale hyperparameter recovery {#lengthscale}

(ref:lengthscale-recovery) Figure caption.
(ref:lengthscale-recovery) The lengthscale posterior mean and 95% credible interval obtained using the centroid kernel model on integrated kernel data for the first 40 simulation replicates on each geometry. The true lengthscale, and lengthscale obtained using the heuristic method of @best1999bayesian are shown as dashed lines.

```{r lengthscale-recovery, fig.cap="(ref:lengthscale-recovery)"}
path <- paste0("resources/beyond-borders/", resource_version, "/depends/lengthscale-recovery.png")
Expand Down
98 changes: 49 additions & 49 deletions docs/a-model-for-risk-group-proportions.html
View file
Original file line number Diff line number Diff line change
Expand Up @@ -495,12 +495,12 @@ <h2><span class="header-section-number">B.1</span> The Global AIDS Strategy<a hr
</div>
<div id="household-survey-data-1" class="section level2 hasAnchor" number="9.2">
<h2><span class="header-section-number">B.2</span> Household survey data<a href="a-model-for-risk-group-proportions.html#household-survey-data-1" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<div id="icefngidqb" style="padding-left:0px;padding-right:0px;padding-top:10px;padding-bottom:10px;overflow-x:auto;overflow-y:auto;width:auto;height:auto;">
<div id="qtkvhcjnqe" style="padding-left:0px;padding-right:0px;padding-top:10px;padding-bottom:10px;overflow-x:auto;overflow-y:auto;width:auto;height:auto;">
<style>html {
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, 'Helvetica Neue', 'Fira Sans', 'Droid Sans', Arial, sans-serif;
}

#icefngidqb .gt_table {
#qtkvhcjnqe .gt_table {
display: table;
border-collapse: collapse;
margin-left: auto;
Expand All @@ -525,7 +525,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-left-color: #D3D3D3;
}

#icefngidqb .gt_heading {
#qtkvhcjnqe .gt_heading {
background-color: #FFFFFF;
text-align: center;
border-bottom-color: #FFFFFF;
Expand All @@ -537,12 +537,12 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-right-color: #D3D3D3;
}

#icefngidqb .gt_caption {
#qtkvhcjnqe .gt_caption {
padding-top: 4px;
padding-bottom: 4px;
}

#icefngidqb .gt_title {
#qtkvhcjnqe .gt_title {
color: #333333;
font-size: 125%;
font-weight: initial;
Expand All @@ -554,7 +554,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-bottom-width: 0;
}

#icefngidqb .gt_subtitle {
#qtkvhcjnqe .gt_subtitle {
color: #333333;
font-size: 85%;
font-weight: initial;
Expand All @@ -566,13 +566,13 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-top-width: 0;
}

#icefngidqb .gt_bottom_border {
#qtkvhcjnqe .gt_bottom_border {
border-bottom-style: solid;
border-bottom-width: 2px;
border-bottom-color: #D3D3D3;
}

#icefngidqb .gt_col_headings {
#qtkvhcjnqe .gt_col_headings {
border-top-style: solid;
border-top-width: 2px;
border-top-color: #D3D3D3;
Expand All @@ -587,7 +587,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-right-color: #D3D3D3;
}

#icefngidqb .gt_col_heading {
#qtkvhcjnqe .gt_col_heading {
color: #333333;
background-color: #FFFFFF;
font-size: 100%;
Expand All @@ -607,7 +607,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
overflow-x: hidden;
}

#icefngidqb .gt_column_spanner_outer {
#qtkvhcjnqe .gt_column_spanner_outer {
color: #333333;
background-color: #FFFFFF;
font-size: 100%;
Expand All @@ -619,15 +619,15 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
padding-right: 4px;
}

#icefngidqb .gt_column_spanner_outer:first-child {
#qtkvhcjnqe .gt_column_spanner_outer:first-child {
padding-left: 0;
}

#icefngidqb .gt_column_spanner_outer:last-child {
#qtkvhcjnqe .gt_column_spanner_outer:last-child {
padding-right: 0;
}

#icefngidqb .gt_column_spanner {
#qtkvhcjnqe .gt_column_spanner {
border-bottom-style: solid;
border-bottom-width: 2px;
border-bottom-color: #D3D3D3;
Expand All @@ -639,7 +639,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
width: 100%;
}

#icefngidqb .gt_group_heading {
#qtkvhcjnqe .gt_group_heading {
padding-top: 8px;
padding-bottom: 8px;
padding-left: 5px;
Expand All @@ -665,7 +665,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
text-align: left;
}

#icefngidqb .gt_empty_group_heading {
#qtkvhcjnqe .gt_empty_group_heading {
padding: 0.5px;
color: #333333;
background-color: #FFFFFF;
Expand All @@ -680,15 +680,15 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
vertical-align: middle;
}

#icefngidqb .gt_from_md > :first-child {
#qtkvhcjnqe .gt_from_md > :first-child {
margin-top: 0;
}

#icefngidqb .gt_from_md > :last-child {
#qtkvhcjnqe .gt_from_md > :last-child {
margin-bottom: 0;
}

#icefngidqb .gt_row {
#qtkvhcjnqe .gt_row {
padding-top: 8px;
padding-bottom: 8px;
padding-left: 5px;
Expand All @@ -707,7 +707,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
overflow-x: hidden;
}

#icefngidqb .gt_stub {
#qtkvhcjnqe .gt_stub {
color: #333333;
background-color: #FFFFFF;
font-size: 100%;
Expand All @@ -720,7 +720,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
padding-right: 5px;
}

#icefngidqb .gt_stub_row_group {
#qtkvhcjnqe .gt_stub_row_group {
color: #333333;
background-color: #FFFFFF;
font-size: 100%;
Expand All @@ -734,11 +734,11 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
vertical-align: top;
}

#icefngidqb .gt_row_group_first td {
#qtkvhcjnqe .gt_row_group_first td {
border-top-width: 2px;
}

#icefngidqb .gt_summary_row {
#qtkvhcjnqe .gt_summary_row {
color: #333333;
background-color: #FFFFFF;
text-transform: inherit;
Expand All @@ -748,16 +748,16 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
padding-right: 5px;
}

#icefngidqb .gt_first_summary_row {
#qtkvhcjnqe .gt_first_summary_row {
border-top-style: solid;
border-top-color: #D3D3D3;
}

#icefngidqb .gt_first_summary_row.thick {
#qtkvhcjnqe .gt_first_summary_row.thick {
border-top-width: 2px;
}

#icefngidqb .gt_last_summary_row {
#qtkvhcjnqe .gt_last_summary_row {
padding-top: 8px;
padding-bottom: 8px;
padding-left: 5px;
Expand All @@ -767,7 +767,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-bottom-color: #D3D3D3;
}

#icefngidqb .gt_grand_summary_row {
#qtkvhcjnqe .gt_grand_summary_row {
color: #333333;
background-color: #FFFFFF;
text-transform: inherit;
Expand All @@ -777,7 +777,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
padding-right: 5px;
}

#icefngidqb .gt_first_grand_summary_row {
#qtkvhcjnqe .gt_first_grand_summary_row {
padding-top: 8px;
padding-bottom: 8px;
padding-left: 5px;
Expand All @@ -787,11 +787,11 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-top-color: #D3D3D3;
}

#icefngidqb .gt_striped {
#qtkvhcjnqe .gt_striped {
background-color: rgba(128, 128, 128, 0.05);
}

#icefngidqb .gt_table_body {
#qtkvhcjnqe .gt_table_body {
border-top-style: solid;
border-top-width: 2px;
border-top-color: #D3D3D3;
Expand All @@ -800,7 +800,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-bottom-color: #D3D3D3;
}

#icefngidqb .gt_footnotes {
#qtkvhcjnqe .gt_footnotes {
color: #333333;
background-color: #FFFFFF;
border-bottom-style: none;
Expand All @@ -814,7 +814,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-right-color: #D3D3D3;
}

#icefngidqb .gt_footnote {
#qtkvhcjnqe .gt_footnote {
margin: 0px;
font-size: 90%;
padding-left: 4px;
Expand All @@ -823,7 +823,7 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
padding-right: 5px;
}

#icefngidqb .gt_sourcenotes {
#qtkvhcjnqe .gt_sourcenotes {
color: #333333;
background-color: #FFFFFF;
border-bottom-style: none;
Expand All @@ -837,72 +837,72 @@ <h2><span class="header-section-number">B.2</span> Household survey data<a href=
border-right-color: #D3D3D3;
}

#icefngidqb .gt_sourcenote {
#qtkvhcjnqe .gt_sourcenote {
font-size: 90%;
padding-top: 4px;
padding-bottom: 4px;
padding-left: 5px;
padding-right: 5px;
}

#icefngidqb .gt_left {
#qtkvhcjnqe .gt_left {
text-align: left;
}

#icefngidqb .gt_center {
#qtkvhcjnqe .gt_center {
text-align: center;
}

#icefngidqb .gt_right {
#qtkvhcjnqe .gt_right {
text-align: right;
font-variant-numeric: tabular-nums;
}

#icefngidqb .gt_font_normal {
#qtkvhcjnqe .gt_font_normal {
font-weight: normal;
}

#icefngidqb .gt_font_bold {
#qtkvhcjnqe .gt_font_bold {
font-weight: bold;
}

#icefngidqb .gt_font_italic {
#qtkvhcjnqe .gt_font_italic {
font-style: italic;
}

#icefngidqb .gt_super {
#qtkvhcjnqe .gt_super {
font-size: 65%;
}

#icefngidqb .gt_footnote_marks {
#qtkvhcjnqe .gt_footnote_marks {
font-style: italic;
font-weight: normal;
font-size: 75%;
vertical-align: 0.4em;
}

#icefngidqb .gt_asterisk {
#qtkvhcjnqe .gt_asterisk {
font-size: 100%;
vertical-align: 0;
}

#icefngidqb .gt_indent_1 {
#qtkvhcjnqe .gt_indent_1 {
text-indent: 5px;
}

#icefngidqb .gt_indent_2 {
#qtkvhcjnqe .gt_indent_2 {
text-indent: 10px;
}

#icefngidqb .gt_indent_3 {
#qtkvhcjnqe .gt_indent_3 {
text-indent: 15px;
}

#icefngidqb .gt_indent_4 {
#qtkvhcjnqe .gt_indent_4 {
text-indent: 20px;
}

#icefngidqb .gt_indent_5 {
#qtkvhcjnqe .gt_indent_5 {
text-indent: 25px;
}
</style>
Expand Down Expand Up @@ -1647,14 +1647,14 @@ <h2><span class="header-section-number">B.4</span> Survey questions and risk gro
<div id="additional-figures" class="section level2 hasAnchor" number="9.5">
<h2><span class="header-section-number">B.5</span> Additional figures<a href="a-model-for-risk-group-proportions.html#additional-figures" class="anchor-section" aria-label="Anchor link to header"></a></h2>

<div class="figure" style="text-align: center"><span style="display:block;" id="fig:aaa-variance-proportions"></span>
<div class="figure" style="text-align: centre"><span style="display:block;" id="fig:aaa-variance-proportions"></span>
<img src="resources/multi-agyw/20231219-192637-4244d933/depends/aaa-variance-proportions.png" alt="The proportion of posterior variance explained by each random effect, calculated as a ratio of the random effect variance posterior mean to the sum of all random effect variance posterior means. To allow calculation of this metric by country, the model was run for each country individually." width="95%" />
<p class="caption">
Figure B.1: The proportion of posterior variance explained by each random effect, calculated as a ratio of the random effect variance posterior mean to the sum of all random effect variance posterior means. To allow calculation of this metric by country, the model was run for each country individually.
</p>
</div>

<div class="figure" style="text-align: center"><span style="display:block;" id="fig:age-variation"></span>
<div class="figure" style="text-align: centre"><span style="display:block;" id="fig:age-variation"></span>
<img src="resources/multi-agyw/20231219-192637-4244d933/depends/age-variation.png" alt="For the 20-24 and 25-29 age groups, the proportion of AGYW in the one cohabiting partner and non-regular or multiple partner(s) risk groups was bimodal." width="95%" />
<p class="caption">
Figure B.2: For the 20-24 and 25-29 age groups, the proportion of AGYW in the one cohabiting partner and non-regular or multiple partner(s) risk groups was bimodal.
Expand Down
Loading

0 comments on commit e7b1cc9

Please sign in to comment.