Type:	Package
Title:	Graph Prediction from a Graph Time Series
Version:	0.1.3
Maintainer:	Sevvandi Kandanaarachchi <sevvandik@gmail.com>
Description:	Predicting the structure of a graph including new nodes and edges using a time series of graphs. Flux balance analysis, a linear and integer programming technique used in biochemistry is used with time series prediction methods to predict the graph structure at a future time point Kandanaarachchi (2025) <doi:10.48550/arXiv.2507.05806>.
License:	GPL (≥ 3)
Encoding:	UTF-8
LazyData:	true
RoxygenNote:	7.3.2
Imports:	dplyr, fable, fabletools, forecast, future, igraph, lpSolve, Matrix, rlang, stats, tibble, feasts, tsibble
Suggests:	knitr, nnet, rmarkdown, urca
URL:	https://sevvandi.github.io/netseer/
VignetteBuilder:	knitr
Depends:	R (≥ 3.5)
NeedsCompilation:	no
Packaged:	2026-03-23 06:49:47 UTC; kan092
Author:	Sevvandi Kandanaarachchi [aut, cre], Stefan Westerlund [aut]
Repository:	CRAN
Date/Publication:	2026-03-23 08:00:02 UTC

Generates a bigger graph using exponential growth.

Description

Generates a bigger graph using parameters for node and edge growth. If a sequence of graphs are created, the number of nodes in this sequence would exponentially increase.

Usage

generate_graph_exp(
  gr = NULL,
  del_edge = 0.1,
  new_nodes = 0.1,
  edge_increase = 0.1
)

Arguments

Value

A graph.

Examples

set.seed(1)
gr <- generate_graph_exp()
gr

Generates a bigger graph by linear growth.

Description

Generates a bigger graph using parameters for node and edge growth. If a sequence of graphs are created, the number of nodes would linearly increase.

Usage

generate_graph_linear(
  gr = NULL,
  del_edge = 1,
  new_nodes = 1,
  edge_increase = 1,
  edges_per_new_node = 3
)

Arguments

Value

A graph.

Examples

set.seed(1)
gr <- generate_graph_linear()
gr

Loads graphs to memory using a desired method.

Description

This function loads graphs from the file system into the R environment. There are two loading options: Loading all files from a directory. Loading individual graphs.

Usage

load_graphs(use_directory = NULL, use_list = NULL, format)

Arguments

Value

A list of graphs in igraph format.

Examples

## Not run: 
graph_dir_list <- list()
path1 <- normalizePath("graph1.gml")
path2 <- normalizePath("graph2.gml")
graph_dir <- append(graph_dir_list, path1 )
graph_dir <- append(graph_dir_list, path2 )

grlist <- read_graph_list(graph_dir_list, "gml")
grlist

## End(Not run)

Loads all graphs in a directory.

Description

This function loads all graphs from a directory to a list

Usage

load_graphs_dir(path_to_graphs, format)

Arguments

Value

A list of graphs in igraph format.

Examples

## Not run: 
path_to_graphs <- normalizePath("./path/to/graphs/")
grlist <- read_graph_list(path_to_graphs, "gml")
grlist

## End(Not run)

Loads graphs using a list of individual paths to each graph.

Description

This function adds graphs to a list by reading in a list of paths to individual graphs.

Usage

load_graphs_list(graph_dir_list, format)

Arguments

Value

A list of graphs in igraph format.

Examples

## Not run: 
graph_dir_list <- list()
path1 <- normalizePath("graph1.gml")
path2 <- normalizePath("graph2.gml")
graph_dir <- append(graph_dir_list, path1 )
graph_dir <- append(graph_dir_list, path2 )

grlist <- read_graph_list(graph_dir_list, "gml")
grlist

## End(Not run)

Gives an error measurement for predicted graphs

Description

This function compares the predicted graph with the actual and comptues the node and edge error as a proportion

Usage

measure_error(actual, predicted)

Arguments

Value

The node error and edge error as a proportion.

Examples

data(syngraphs)
# Taking the 20th graph as the actual and the 19th graph as predicted.
measure_error(syngraphs[[20]], syngraphs[[19]])

Predicts a graph from a time series of graphs.

Description

This function predicts the graph at a future time step using a time series of graphs.

Usage

predict_graph(
  graphlist,
  formulation = 2,
  conf_level1 = NULL,
  conf_level2 = 90,
  dense_opt = 2,
  weights_opt = 8,
  weights_param = 0.001,
  h = 1
)

Arguments

Value

A list of predicted graphs. If conf_level1 is not NULL, then 3 graphs are returned one with the mean number of predicted nodes and the other 2 with the number of nodes equal to the lower and upper bound values of prediction. If If conf_level1 is NULL, only the mean predicted graph is returned.

Examples

set.seed(2024)
edge_increase_val <- new_nodes_val <- del_edge_val <- 0.1
graphlist <- list()
graphlist[[1]] <- gr <-  igraph::sample_pa(5, directed = FALSE)
for(i in 2:15){
  gr <-  generate_graph_exp(gr,
                        del_edge = del_edge_val,
                        new_nodes = new_nodes_val,
                        edge_increase = edge_increase_val )
  graphlist[[i]] <- gr
}
grpred <- predict_graph(graphlist[1:15], conf_level2 = 90, weights_opt = 6)
grpred
 

Saves either a single graph or list of graphs to disk.

Description

This function saves a single graph or list of graphs to a specified location on the file system in a specified format.

Usage

save_graphs(graph, file_path, filetype = ".gml", format)

Arguments

Value

A list of graphs in igraph format.

Examples

## Not run: 
library(igraph)
sample_graph  <- igraph::graph_from_literal(A-B, B-C)
path <- "/path/to/save/to/"
save_graphs(sample_graph, path, ".gml", "gml")

## End(Not run)

A dataset containing synthetic graphs

Description

This dataset contains a list of synthetic igraph objects

Usage

syngraphs

Format

A list of 20 igraph graphs

Examples

data(syngraphs)
syngraphs[[1]]