netreducing: calculate variance in network level metrics caused by sample...

In hemprichbennett/netReducer: Analysing the impact of sample size and species composition in network analysis

Description

Acts as a wrapper for the shrink function. IMPORTANT: the columns of your matrix must be named speciesname-samplename, with a hyphen separating them. E.g. species1-sample1, species2-sampleb, etc etc. The hyphen is used to split the names and so is (currently) non-negotiable.

Usage

netreducing(input, input_type, n_iterations = 100, min_nodes, metric_chosen,
  type_chosen, level = NA, collapse_cols = T)

Arguments

input	Your input object
input_type	The type of your input object. Can be either 'matrix' or 'list'
n_iterations	The number of iterations you wish to perform per number of nodes selected
min_nodes	The minimum number of columns to be allowed in the matrix tested. For example, if you input a matrix with ten columns and set a min_nodes of 5, the function will test your network with 5, 6, 7, 8, and 9 nodes.
metric_chosen	The metric you would like to analyse. Can be any value allowed by bipartite for the networklevel or specieslevel functions, or if 'modularity' is chosen, it will use computemodules to calculate modularity
type_chosen	The type of metric desired. Can either be 'network' for a networklevel metric, 'species' for a specieslevel metric, or 'modularity' for modularity
level	The network level desired to be analysed.
collapse_cols	Should the columns be collapsed by species? I.e. is this a species-based or individual-based analysis

Value

Produces a dataframe showing the changes in a given metric when varying the number of nodes in your matrix

Examples

pending

hemprichbennett/netReducer documentation built on Aug. 11, 2019, 6:48 a.m.