generateC: Generate a matrix of pairwise wins

In StatRank: Statistical Rank Aggregation: Inference, Evaluation, and Visualization

Description

This function takes in data that has been broken up into pair format. The user is given a matrix C, where element C[i, j] represents (if normalized is FALSE) exactly how many times alternative i has beaten alternative j (if normalized is TRUE) the observed probability that alternative i beats j

Usage

generateC(Data.pairs, m, weighted = FALSE, prior = 0, normalized = TRUE)

Arguments

Data.pairs	the data broken up into pairs
m	the tot al number of alternatives
weighted	whether or not this generateC should use the third column of Data.pairs as the weights
prior	the initial "fake data" that you want to include in C. A prior of 1 would mean that you initially "observe" that all alternatives beat all other alternatives exactly once.
normalized	if TRUE, then normalizes entries to probabilities

Value

a Count matrix of how many times alternative i has beat alternative j

Examples

data(Data.Test)
Data.Test.pairs <- Breaking(Data.Test, "full")
generateC(Data.Test.pairs, 5)

StatRank documentation built on May 1, 2019, 8:22 p.m.