corankco 1.0.0

Kemeny-Young method for rank aggregation of incomplete rankings with ties

This package implements methods for rank aggregation of incomplete rankings with ties

Installation

Install from PyPI:

pip3 install --user corankco

Example usage

>>> from corankco.dataset import Dataset
>>> from corankco.scoringscheme import ScoringScheme
>>> from corankco.kemrankagg import KemRankAgg
>>> from corankco.algorithms.enumeration import Algorithm
>>>
>>> d = Dataset([
...               [[1], [2, 3]],
...               [[3, 1], [4]],
...               [[1], [5], [3, 2]]
...              ])
>>> print(d.description())
Dataset description:
        elements:5
        rankings:3
        complete:False
        with ties: True
        rankings:
                r1 = [[1], [2, 3]]
                r2 = [[3, 1], [4]]
                r3 = [[1], [5], [3, 2]]

>>> # Generates default scoring scheme
>>> sc = ScoringScheme()

>>> # Consensus computation with an exact algorithm
>>> consensus = KemRankAgg.compute_consensus(d, sc, Algorithm.Exact)

>>> print(consensus.description())
Consensus description:
        computed by:Exact algorithm ILP Cplex
        necessarily optimal:True
        kemeny score:6.0
        consensus:
                c1 = [[1], [2, 3], [4], [5]]
                c2 = [[1], [2, 3], [5], [4]]

>>> # Consensus computation with an heuristic
consensus = KemRankAgg.compute_consensus(d, sc, Algorithm.ParCons)

>>> print(consensus.description())
Consensus description:
        necessarily optimal:True
        computed by:ParCons, uses  "BioConsert with input rankings as starters" on groups of size >  80
        weak partitioning (at least one optimal solution)[{1}, {2, 3}, {5}, {4}]
        kemeny score:8.0
        consensus:
                c1 = [[1], [2, 3], [5], [4]]

>>> # example of computing score ('distance') between two ranking
>>> r1 = [[1], [2], [3, 4]]
>>> r2 = [[3], [2]]
>>> print(KemenyScoreFactory.score_between_rankings(r1, r2, sc))
5.