lenskit.metrics.ranking.GeometricRankWeight
===========================================

.. py:class:: lenskit.metrics.ranking.GeometricRankWeight(patience = 0.85)
   :canonical: lenskit.metrics.ranking._weighting.GeometricRankWeight

   Bases: :py:obj:`RankWeight`


   Geometric cascade weighting for result ranks.

   This is the ranking model used by RBP :cite:p:`rbp`.

   For patience :math:`p`, the discount is given by :math:`p^{k-1}`.  The sum
   of this infinite series is :math:`\frac{1}{1 - p}`.

   :param patience: The patience parameter :math:`p`.

   :Stability: Caller


   .. py:attribute:: patience
      :type:  float


   .. py:method:: weight(ranks)

      Compute the discount for the specified ranks.

      Ranks must start with 1.



   .. py:method:: log_weight(ranks)

      Compute the (natural) log of the discount for the specified ranks.

      Ranks must start with 1.



   .. py:method:: series_sum()

      Get the sum of the infinite series of this discount function, if known.
      Some metrics (e.g. :func:`~lenskit.metrics.RBP`) will use this to
      normalize their measurements.