mfm_standard: Standard version of the MFM algorithm

In AlexisDerumigny/MFunctMatching: Subsample Selection by Multifunctional Matching

Description

This function uses the MFM algorithm with predefined functionals, which are:

• the mean and standard deviation for each numeric or integer variable

• Kendall's tau for each pair of numeric / integer variable;

• the empirical proportion for each unique possibility of each variable of type character or factor.

Usage

mfm_standard(
  data,
  coordX = NULL,
  coordY = NULL,
  minDistance = 0,
  nDraws,
  sizeSample,
  methodNormalization = "ecdf"
)

Arguments

data	the dataframe to use
coordX	the vector of X-coordinates
coordY	the vector of Y-coordinates
minDistance	minimum distance between two observations of the selected sample
nDraws	the number of random samples to generate from the population
sizeSample	the size of the sample
methodNormalization	the method used for normalization

Examples

df = data.frame(X1 = rnorm(2000), X2 = rnorm(2000),
                X3 = c(rep("A",500), rep("B", 500), rep("C", 1000)))
mfm_standard(data = df, nDraws = 500, sizeSample = 10,
             coordX = rnorm(2000), coordY = rnorm(2000))

# agricultural example
df = agridat::gartner.corn
mfm_standard(data = df[,c("mass", "moist", "elev")],
             nDraws = 500, sizeSample = 10,
             coordX = df$long, coordY = df$lat)

AlexisDerumigny/MFunctMatching documentation built on Dec. 31, 2020, 9:47 a.m.