Sha256: 02fb304f6bf69d49ac70ce9449d9f280a3ff9a3218126ab63145c0eb800b8de8
Contents?: true
Size: 918 Bytes
Versions: 29
Compression:
Stored size: 918 Bytes
Contents
# frozen_string_literal: true
require 'rumale/base/evaluator'
module Rumale
# This module consists of the classes for model evaluation.
module EvaluationMeasure
# Accuracy is a class that calculates the accuracy of classifier from the predicted labels.
#
# @example
# evaluator = Rumale::EvaluationMeasure::Accuracy.new
# puts evaluator.score(ground_truth, predicted)
class Accuracy
include Base::Evaluator
# Calculate mean accuracy.
#
# @param y_true [Numo::Int32] (shape: [n_samples]) Ground truth labels.
# @param y_pred [Numo::Int32] (shape: [n_samples]) Predicted labels.
# @return [Float] Mean accuracy
def score(y_true, y_pred)
check_label_array(y_true)
check_label_array(y_pred)
(y_true.to_a.map.with_index { |label, n| label == y_pred[n] ? 1 : 0 }).inject(:+) / y_true.size.to_f
end
end
end
end
Version data entries
29 entries across 29 versions & 1 rubygems