nroSummary {Numero}

Estimate subgroup statistics

Description

Combine data points that reside in districts that belong to a larger region into a subgroup; compare descriptive statistics between subgroups.

Usage

nroSummary(data, districts, regions = NULL, categlim = 8, capacity = 10)

Arguments

data

A vector of named M elements or an M x N matrix of data values with row names.

districts

An integer vector of M named elements that indicate the best match out of K districts for each row name in the data matrix, please see nroMatch for an example.

regions

An vector of K elements or a data frame of K rows that defines if a district belongs to a larger region (i.e. a subgroup), see details.

categlim

The threshold for the number of unique values before a variable is considered continuous.

capacity

Maximum number of subgroups to compare.

Details

If defined, the region vector should have K elements where K is the total number of map districts.

The region input can also be a data frame of K rows where the column REGION will be used for assigning district to regions, and REGION.label will be used as the character label as seen on the map, see the output from nroPlot() for an example.

Districts and data points are connected by comparing element names in districts and names or row names of data.

Districts and regions are connected by comparing element values in districts and names or row names of regions.

If the region vector is empty, each district is automatically assigned to its own region.

Safeguards are in place to prevent crashes from empty categories; this reduces statistical power slightly when numbers are small.

Value

A data frame of summary statistics that contains a row for every combination of subgroups and variables. The chi-squared test is used for comparisons with respect to categorical variables, and rank-regulated t-test and ANOVA are applied to continuous variables. Region labels for each row are stored in the attribute 'labels' and a list that contains the subsets of rows in each region is stored in the attribute 'subgroups'.

Examples

# Import data.
fname <- system.file("extdata", "finndiane.txt", package = "Numero")
dataset <- read.delim(file = fname)

# Prepare training data.
trvars <- c("CHOL", "HDL2C", "TG", "CREAT", "uALB")
trdata <- scale.default(dataset[,trvars])

# K-means clustering.
km <- nroKmeans(data = trdata)

# Self-organizing map.
sm <- nroKohonen(seeds = km)
sm <- nroTrain(map = sm, data = trdata)

# Assign data points into districts.
matches <- nroMatch(centroids = sm, data = trdata)

# Calculate district averages for urinary albumin.
plane <- nroAggregate(topology = sm, districts = matches,
                      data = dataset$uALB)
plane <- as.vector(plane)

# Assign subgroups based on urinary albumin.
regns <- rep("HighAlb", length.out=length(plane))
regns[which(plane < quantile(plane, 0.67))] <- "MiddleAlb"
regns[which(plane < quantile(plane, 0.33))] <- "LowAlb"

# Add label info and make a data frame.
regns <- data.frame(REGION=regns, REGION.label="",
    stringsAsFactors=FALSE)
regns[which(regns$REGION == "HighAlb"),"REGION.label"] <- "H"
regns[which(regns$REGION == "MiddleAlb"),"REGION.label"] <- "M"
regns[which(regns$REGION == "LowAlb"),"REGION.label"] <- "L"

# Calculate summary statistics.
st <- nroSummary(data = dataset, districts = matches, regions = regns)
print(st[,c("VARIABLE","SUBGROUP","MEAN","P.chisq","P.t","P.anova")])