Class HeuristicEmbedednessReranker<U>

Type Parameters:
U - type of the users
All Implemented Interfaces:
GlobalReranker<U,​U>

public class HeuristicEmbedednessReranker<U>
extends AbstractHeuristicNeighborOverlapReranker<U>
Swap reranker for optimizing the embeddedness of the graph. It uses heuristics to optimize the execution times.
  • Constructor Details

    • HeuristicEmbedednessReranker

      public HeuristicEmbedednessReranker​(double lambda, int cutOff, java.util.function.Supplier<Normalizer<U>> norm, Graph<U> graph, int mode)
      Constructor
      Parameters:
      cutOff - the definitive length of the recommendation rankings.
      lambda - trade-off between the average embeddedness and the original score
      norm - the normalization scheme.
      graph - the original graph
      mode - the execution mode: 1) Embededness is corrected any time a swap is done. 2) Embededness is corrected every time a user has finished its reranking. 3) Embededness is never corrected (only use the heuristic)