From eac1bc09931e19ba61f9fc2a58a6e4a34886ec00 Mon Sep 17 00:00:00 2001 From: midinas <midinas@users.noreply.github.com> Date: Thu, 29 Jun 2017 13:06:08 -0700 Subject: [PATCH] Fix in documentation doc/modules/clustering.rst (#9243) Added missing minus sign in entropy formulas that explain mutual information scoring. --- doc/modules/clustering.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/doc/modules/clustering.rst b/doc/modules/clustering.rst index 0167fef69f..4dda5559b0 100644 --- a/doc/modules/clustering.rst +++ b/doc/modules/clustering.rst @@ -1119,12 +1119,12 @@ Mathematical formulation Assume two label assignments (of the same N objects), :math:`U` and :math:`V`. Their entropy is the amount of uncertainty for a partition set, defined by: -.. math:: H(U) = \sum_{i=1}^{|U|}P(i)\log(P(i)) +.. math:: H(U) = - \sum_{i=1}^{|U|}P(i)\log(P(i)) where :math:`P(i) = |U_i| / N` is the probability that an object picked at random from :math:`U` falls into class :math:`U_i`. Likewise for :math:`V`: -.. math:: H(V) = \sum_{j=1}^{|V|}P'(j)\log(P'(j)) +.. math:: H(V) = - \sum_{j=1}^{|V|}P'(j)\log(P'(j)) With :math:`P'(j) = |V_j| / N`. The mutual information (MI) between :math:`U` and :math:`V` is calculated by: -- GitLab