Breaking: Breaks full or partial orderings into pairwise comparisons

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

Description

Given full or partial orderings, this function will generate pairwise comparison Options 1. full - All available pairwise comparisons. This is used for partial rank data where the ranked objects are a random subset of all objects 2. adjacent - Only adjacent pairwise breakings 3. top - also takes in k, will break within top k and will also generate pairwise comparisons comparing the top k with the rest of the data 4. top.partial - This is used for partial rank data where the ranked alternatives are preferred over the non-ranked alternatives

Usage

Breaking(Data, method, k = NULL)

Arguments

Data	data in either full or partial ranking format
method	- can be full, adjacent, top or top.partial
k	This applies to the top method, choose which top k to focus on

Value

Pairwise breakings, where the three columns are winner, loser and rank distance (latter used for Zemel)

Examples

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

Example output

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