2022-06-15-singleRcapture-showcase.Rmd

---
title: "R Notebook"
output: html_notebook
---

## Outline

1. general idea behind the package
2. presentation of the package and its functionalities
  + main function `estimate_popsize`
  + diagnostics (residuals, leave-one-out diagnostics)
  + testing  `marginalFreq`
  + datasets `carcassubmission`, `netherlandsimmigrant`
3. workplan
4. discussion

## Installation

Installation of the package from github [working branch]

```{r}
remotes::install_github("ncn-foreigners/singleRcapture@working")
```

Loading the package

```{r}
library(singleRcapture)
```

Main functions:

+ `estimate_popsize` -- declaration of formulae, type of model (i.e. distribution), estimation method, variance estimation and many other, under the hood:
  + `estimate_popsize.fit` -- function to fit the model (in matrix notation)
  + `populationEstimate` -- function to estimate the population size given input in `estimate_popsize`
+ `marginalFreq` -- function to verify esti


```{r}
m1 <- estimate_popsize(formula = capture ~ gender + age + nation, 
                       model = "ztpoisson", 
                       method = "mle", 
                       pop.var = "analytic", 
                       data = netherlandsimmigrant)
```

Basic output

```{r}
m1
```

Detailed output

```{r}
summary(m1)
```
Basic, stats functions

```{r}
table(resid(m1))
```

Multiple models

```{r}
selected_models <- c("ztpoisson", "zelterman", "chao", "ztgeom")

multiple_models <- lapply(selected_models, 
                          function(x) {
                            estimate_popsize(formula = capture ~ gender + age + nation, 
                                             model = x, 
                                             method = "mle", 
                                             pop.var = "analytic",
                                             data = netherlandsimmigrant)
                            })

pop_sizes <- lapply(multiple_models, "[[", "populationSize")
```

```{r}
data.frame(x = selected_models,
           N_hat = sapply(pop_sizes, "[[", "pointEstimate"))
```

```{r}
data.frame(x = selected_models,
           N_hat = do.call("rbind", lapply(pop_sizes, "[[", "confidenceInterval")))
```