Class PlackettLuceModel

java.lang.Object
com.knezevic.edaf.v3.models.permutation.PlackettLuceModel
All Implemented Interfaces:
Model<PermutationVector>

public final class PlackettLuceModel extends Object implements Model<PermutationVector>
Plackett-Luce model for permutations.

Each item i has non-negative weight w_i. A permutation is generated sequentially by sampling one item at a time from the remaining set:

   P(i chosen at step k) = w_i / Σ_{j in remaining} w_j
 
Weights are estimated from elite position scores, then normalized.