Class: Rumale::KernelMachine::KernelRidge

Inherits:
Base::Estimator show all
Includes:
Base::Regressor
Defined in:
rumale-kernel_machine/lib/rumale/kernel_machine/kernel_ridge.rb

Overview

KernelRidge is a class that implements kernel ridge regression.

Examples:

require 'numo/linalg/autoloader'
require 'rumale/pairwise_metric'
require 'rumale/kernel_machine/kernel_ridge'

kernel_mat_train = Rumale::PairwiseMetric::rbf_kernel(training_samples)
kridge = Rumale::KernelMachine::KernelRidge.new(reg_param: 1.0)
kridge.fit(kernel_mat_train, traininig_values)

kernel_mat_test = Rumale::PairwiseMetric::rbf_kernel(test_samples, training_samples)
results = kridge.predict(kernel_mat_test)

Instance Attribute Summary collapse

Attributes inherited from Base::Estimator

#params

Instance Method Summary collapse

Methods included from Base::Regressor

#score

Constructor Details

#initialize(reg_param: 1.0) ⇒ KernelRidge

Create a new regressor with kernel ridge regression.

Parameters:

  • reg_param (Float/Numo::DFloat) (defaults to: 1.0) —

    The regularization parameter.



32
33
34
35
36
37
 
# File 'rumale-kernel_machine/lib/rumale/kernel_machine/kernel_ridge.rb', line 32

def initialize(reg_param: 1.0)
  super()
  @params = {
    reg_param: reg_param
  }
end

Instance Attribute Details

#weight_vec ⇒ Numo::DFloat (readonly)

Return the weight vector.

Returns:

  • (Numo::DFloat) —

    (shape: [n_training_sample, n_outputs])



27
28
29
 
# File 'rumale-kernel_machine/lib/rumale/kernel_machine/kernel_ridge.rb', line 27

def weight_vec
  @weight_vec
end

Instance Method Details

#fit(x, y) ⇒ KernelRidge

Fit the model with given training data.

Parameters:

  • x (Numo::DFloat) —

    (shape: [n_training_samples, n_training_samples]) The kernel matrix of the training data to be used for fitting the model.

  • y (Numo::DFloat) —

    (shape: [n_samples, n_outputs]) The taget values to be used for fitting the model.

Returns:

Raises:

  • (ArgumentError) 


45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
 
# File 'rumale-kernel_machine/lib/rumale/kernel_machine/kernel_ridge.rb', line 45

def fit(x, y)
  x = ::Rumale::Validation.check_convert_sample_array(x)
  y = ::Rumale::Validation.check_convert_target_value_array(y)
  ::Rumale::Validation.check_sample_size(x, y)
  raise ArgumentError, 'Expect the kernel matrix of training data to be square.' unless x.shape[0] == x.shape[1]
  raise 'KernelRidge#fit requires Numo::Linalg but that is not loaded.' unless enable_linalg?(warning: false)

  n_samples = x.shape[0]

  if @params[:reg_param].is_a?(Float)
    reg_kernel_mat = x + Numo::DFloat.eye(n_samples) * @params[:reg_param]
    @weight_vec = Numo::Linalg.solve(reg_kernel_mat, y, driver: 'sym')
  else
    n_outputs = y.shape[1]
    @weight_vec = Numo::DFloat.zeros(n_samples, n_outputs)
    n_outputs.times do |n|
      reg_kernel_mat = x + Numo::DFloat.eye(n_samples) * @params[:reg_param][n]
      @weight_vec[true, n] = Numo::Linalg.solve(reg_kernel_mat, y[true, n], driver: 'sym')
    end
  end

  self
end

#predict(x) ⇒ Numo::DFloat

Predict values for samples.

Parameters:

  • x (Numo::DFloat) —

    (shape: [n_testing_samples, n_training_samples]) The kernel matrix between testing samples and training samples to predict values.

Returns:

  • (Numo::DFloat) —

    (shape: [n_samples, n_outputs]) Predicted values per sample.



74
75
76
77
78
 
# File 'rumale-kernel_machine/lib/rumale/kernel_machine/kernel_ridge.rb', line 74

def predict(x)
  x = ::Rumale::Validation.check_convert_sample_array(x)

  x.dot(@weight_vec)
end