summary.rankComp: Summarizing Ranking Comparison

In giniCI: Gini-Based Composite Indicators

Summarizing Ranking Comparison

Description

Summary method for class "rankComp" and print method for class "summary.rankComp".

Usage

## S3 method for class 'rankComp'
summary(object, n.pick = 10L, n.q = 10L, ...)

## S3 method for class 'summary.rankComp'
print(x, digits = max(3L, getOption("digits") - 3L), ...)

Arguments

object	an object of class "rankComp", usually, an output of a call to rankComp.
n.pick	a positive integer specifying the number of units considered to form the top/bottom based on the alternative index. The default value is 10.
n.q	a positive integer specifying the n-quantiles considered to compute quantile rankings. The default value is 10 (deciles).
...	further arguments passed to or from other methods.
x	an object of class "summary.rankComp", usually, a output of a call to summary.rankComp.
digits	number of significant digits to use when printing.

Details

summary.rankComp provides details on the ranking comparison between the reference and the alternative indices stored in the object of class "rankComp". print.summary.rankComp prints summary information using a smart digit format for the components.

Value

An object of class "summary.rankComp" which is a list of components:

par	a vector storing the values of n.pick and n.q.
n.unit	the number of ranked units (by temporal factors if available).
shift.stats	a data frame with rows presenting the summary statistics of ranking shifts: minimum, first quartile, median, mean, third quartile, and maximum.
asr	a data frame giving the average shift in ranking (ASR) \mathrm{ASR} = \frac{1}{m} \sum^m_{i = 1} | \mathrm{rank}^{\mathrm{alt}}_i - \mathrm{rank}^{\mathrm{ref}}_i |, where m is the number of units considered. The rows present the ASR for all units, and for the top and the bottom units based on the alternative index ranking.
per	a data frame giving the percentage of equal rankings (PER) \mathrm{PER} = 100 \times \frac{1}{m} \sum^m_{i = 1} \{ \mathrm{rank}^{\mathrm{alt}}_i = \mathrm{rank}^{\mathrm{ref}}_i \}, where m is the number of units considered. The rows present the PER for all units, and for the top and the bottom units based on the alternative index ranking.
asq	the average shift in quantile ranking (by temporal factors if available). This value is similar to the ASR for all units, but using the quantile ranking of two indices.

For shift.stats, asr, and per, multiple columns will be generated according to temporal factors if object$time is not NULL.

Author(s)

Viet Duong Nguyen, Chiara Gigliarano, Mariateresa Ciommi

References

Mariani, F., Ciommi, M., and Recchioni, M. C. (2024). Two in One: A New Tool to Combine Two Rankings Based on the Voronoi Diagram. Social Indicators Research, 175, 989–1005.

See Also

rankComp.

Examples

data(bli)

# Goalpost normalization
bli.pol = c("neg", "pos", "pos", "pos", "pos", "neg",
            "pos", "pos", "pos", "neg", "pos")
bli.norm.2014 <- normalize(inds = bli[, 3:13], method = "goalpost",
                           ind.pol = bli.pol, time = bli$YEAR,
                           ref.time = 2014)

# Composite indices
ci.gini <- giniCI(bli.norm.2014, method = "gini",
                  ci.pol = "pos", time = bli$YEAR, ref.time = 2014,
                  only.ci = TRUE)
ci.reci <- giniCI(bli.norm.2014, method = "reci", agg = "geo",
                  ci.pol = "pos", time = bli$YEAR, ref.time = 2014,
                  only.ci = TRUE)

# Ranking comparison
ci.comp <- rankComp(ci.gini, ci.reci, id = bli$COUNTRY, time = bli$YEAR)
print(ci.comp)
summary(ci.comp)

giniCI documentation built on April 3, 2025, 7:35 p.m.