README.Rmd

---
output: github_document
bibliography: reference.bib 
---

<!-- README.md is generated from README.Rmd. Please edit that file -->

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
```

# nethist

<!-- badges: start -->
[![R-CMD-check](https://github.com/EnigmaSong/nethist/actions/workflows/R-CMD-check.yaml/badge.svg)](https://github.com/EnigmaSong/nethist/actions/workflows/R-CMD-check.yaml)
[![Lifecycle: experimental](https://img.shields.io/badge/lifecycle-experimental-orange.svg)](https://lifecycle.r-lib.org/articles/stages.html#experimental)
<!-- badges: end -->

The goal of *nethist* is to estimate graphons by *network histogram* [@wolfe2013Nonparametric; @olhede2014Network]. It also provides extra tools for summary violin plot for networks [@maugis2017Topology] and visualizing network histogram. Use undirected and simple graphs with no self-loops (either igraph or matrix) as inputs for the functions in this package.

To install the package from source, you need C++ and Fortran compilers.

## Installation

You can install the development version of nethist from [GitHub](https://github.com/) with:

``` r
# install.packages("devtools")
devtools::install_github("EnigmaSong/nethist")
```

## Example

Here are basic examples using political blog data set in the package:

### Network histogram 

```{r load_library}
library(nethist)
```

We use *polblog* dataset in the package for our examples. 

```{r summary, echo = FALSE}
A<- igraph::as_adj(polblog, sparse = FALSE)
heatmap(A, Rowv = NA, symm=TRUE, col = c("white","red"))
```

We can estimate a network histogram from the political blog data and plot it.
```{r example}
## Example code using polblog data set
set.seed(42)
hist_polblog <- nethist(polblog, h = 72) #using user-specified bin size.
plot(hist_polblog)
```


### Plotting option

```{r get_reorder, echo = FALSE}
#Get ratio for liberal
ratio_liberal <- table(hist_polblog$cluster, c(rep(1,586),rep(0,638)))
ratio_liberal <- ratio_liberal[,1]/rowSums(ratio_liberal)
liberal_ind <- which(ratio_liberal>=0.5)
conserv_ind <- which(ratio_liberal<0.5)

#Order by cross-party connection
liberal_cp_con <- rowSums(hist_polblog$p_mat[liberal_ind,conserv_ind])
conserv_cp_con <- colSums(hist_polblog$p_mat[liberal_ind,conserv_ind])

ind <- c(rev(liberal_ind[order(liberal_cp_con,decreasing=TRUE)]),
         conserv_ind[order(conserv_cp_con,decreasing=TRUE)],use.names=FALSE);
```

#### heatmap() style

`plot()` provides 2D plot as `heatmap()`.

You can use a user-specified indices for plots. Here is an example:
```{r example2}
print(ind) 
## Users can specify the index order of heatmap
plot(hist_polblog, idx_order = ind)

## Users can specify the color palette
library(RColorBrewer)
plot(hist_polblog,  idx_order = ind, col = brewer.pal(9, "Greys"))
```

You can display the estimated block probabilities by setting `type = prob` and `prob=TRUE`.
```{r example3}
## Users can specify the color palette 
plot(hist_polblog, idx_order = ind, type = "prob", prob= TRUE, prob.col = "blue",
     col = colorRampPalette(colors=c("#FFFFFF","#000000"))(200))
```

### others

There are more types of plots in `nethist` package. 

### Summary violin plot

If you want to check network summary violin plot of the data set:
```{r example4}
#User-specified subsample size.
violin_netsummary(polblog, max_cycle_order = 7, subsample_sizes = 250) 
#Auto-selected subsample size.
violin_netsummary(polblog, max_cycle_order = 7) 
```

## Reference