netCoin (version 2.1.0)

surCoin: Networked coincidences from a data frame.

Description

surCoin produces a network object of coincidences from a data frame converting variables into dichotomies.

Usage

surCoin(data,variables=names(data), commonlabel=NULL,
        dichotomies=NULL, valueDicho=1, metric=NULL, exogenous=NULL,
        weight=NULL, subsample=FALSE, pairwise=FALSE,
        minimum=1, maximum=nrow(data), sort=FALSE, decreasing=TRUE,
        frequency=FALSE, percentages=TRUE,
        procedures="Haberman", criteria="Z", Bonferroni=FALSE,
        support=-Inf, minL=-Inf, maxL=Inf,
        directed=FALSE, diagonal=FALSE, sortL=NULL, decreasingL=TRUE,
        igraph=FALSE, coin=FALSE, dir=NULL, ...)

Value

This function creates a netCoin object (or igraph) and, if stated, a folder in the computer with an HTML document named index.html which contains the produced graph. This file can be directly opened with your browser and sent to a web server to work properly.

Arguments

data

a data frame.

variables

a vector of variables included in the previous data frame.

commonlabel

a vector of variables whose names are to be included in nodes labels.

dichotomies

a vector of dichotomous variables to appear as just one category.

valueDicho

value or values to be selected for dichotomous variables. Default is 1.

metric

a vector of metrics.

exogenous

a vector of variables whose relations amongst them are of no interest. None by default.

weight

a vector of weights. Optimal for data.framed tables.

subsample

retrict the analysis to scenarios with at least one event.

pairwise

Pairwise mode of handling missing values if TRUE. Listwise by default.

minimum

minimum frequency to be considered.

maximum

maximum frequency to be considered.

sort

sort the coincidence matrix according to frequency of events.

decreasing

decreasing or increasing sort of the matrix.

frequency

a logical value true if frequencies are to be shown. Default=FALSE.

percentages

a logical value true if percentages are to be shown. Default=TRUE.

procedures

a vector of statistics of similarity. See below.

criteria

statistic to be use for selection criteria.

Bonferroni

Bonferroni criterium of the signification test.

support

minimum value of the frequency of the coincidence to be edged.

minL

minimum value of the statistic to include the edge in the list.

maxL

maximum value of the statistic to include the edge in the list. By default is +Inf, except if criteria="Z" or criteria="hyp", in which case it is .5. It is recommnended to change it to .05 if data has been sampled.

directed

includes same edges only once.

diagonal

includes auto-links.

sortL

sort the list according to the values of a statistic. See below.

decreasingL

order in a decreasing way.

igraph

Produces an igraph object instead of a netCoin object if TRUE.

coin

Only return the coincidences matrix if TRUE.

dir

a "character" string representing the directory where the web files will be saved.

...

Any netCoin argument.

Author

Modesto Escobar, Department of Sociology and Communication, University of Salamanca. See https://sociocav.usal.es/blog/modesto-escobar/

Details

Possible measures in procedures are

• Frequencies (f), Relative frequencies (x), Conditional frequencies (i), Coincidence degree (cc), Probable degree (cp),

• Expected (e), Confidence interval (con)

• Matching (m), Rogers & Tanimoto (t), Gower (g), Sneath (s), Anderberg (and),

• Jaccard (j), Dice (d), antiDice (a), Ochiai (o), Kulczynski (k),

• Hamann (ham), Yule (y), Pearson (p), odds ratio (od), Rusell (r),

• Haberman (h), Z value of Haberman (z),

• Hypergeometric p greater value (hyp).

• Convert a matrix into an edge list (shape).

References

Escobar, M. and Martinez-Uribe, L. (2020) Network Coincidence Analysis: The netCoin R Package. Journal of Statistical Software, 93, 1-32. tools:::Rd_expr_doi("10.18637/jss.v093.i11").

Examples

# A data frame with two variables Gender and Opinion
frame <- data.frame(Gender=c(rep("Man",3),rep("Woman",3)),
                    Opinion=c("Yes","Yes","No","No","No","Yes"))
surCoin(frame,commonlabel="") # network object

# A data frame with two variables (Gender and Hand) and nodes
input <- data.frame(
  Gender = c("Women", "Men", "Men", "Women", "Women","Men",
             "Men", "Men", "Women", "Women", "Men", "Women"),
  Hand   = c("Right", "Left","Right", "Right", "Right", "Right",
             "Left", "Right", "Right", "Left","Right", "Right"))
nodes <- data.frame(
  name  = c("Gender:Men","Gender:Women", "Hand:Left", "Hand:Right"),
  label = c("Women(50\u25)","Men(50\u25)",
            "Left hand(25\u25)", "Right hand(75\u25)"))
G <- surCoin(input, nodes=nodes, proc=c("h","i"), label="label",
             ltext="i", showArrows=TRUE, maxL=.99)