rumale 0.18.1 → 0.18.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.travis.yml +0 -1
- data/CHANGELOG.md +16 -4
- data/lib/rumale.rb +6 -1
- data/lib/rumale/clustering/dbscan.rb +0 -17
- data/lib/rumale/clustering/gaussian_mixture.rb +0 -21
- data/lib/rumale/clustering/hdbscan.rb +0 -15
- data/lib/rumale/clustering/k_means.rb +0 -17
- data/lib/rumale/clustering/k_medoids.rb +0 -19
- data/lib/rumale/clustering/power_iteration.rb +0 -19
- data/lib/rumale/clustering/single_linkage.rb +0 -17
- data/lib/rumale/clustering/spectral_clustering.rb +0 -17
- data/lib/rumale/evaluation_measure/function.rb +34 -0
- data/lib/rumale/kernel_approximation/rbf.rb +0 -19
- data/lib/rumale/kernel_machine/kernel_pca.rb +0 -21
- data/lib/rumale/kernel_machine/kernel_ridge.rb +0 -15
- data/lib/rumale/kernel_machine/kernel_svc.rb +0 -21
- data/lib/rumale/naive_bayes/base_naive_bayes.rb +47 -0
- data/lib/rumale/naive_bayes/bernoulli_nb.rb +82 -0
- data/lib/rumale/naive_bayes/complement_nb.rb +85 -0
- data/lib/rumale/naive_bayes/gaussian_nb.rb +69 -0
- data/lib/rumale/naive_bayes/multinomial_nb.rb +74 -0
- data/lib/rumale/naive_bayes/negation_nb.rb +71 -0
- data/lib/rumale/nearest_neighbors/k_neighbors_classifier.rb +0 -19
- data/lib/rumale/nearest_neighbors/k_neighbors_regressor.rb +0 -17
- data/lib/rumale/neural_network/adam.rb +0 -19
- data/lib/rumale/preprocessing/bin_discretizer.rb +0 -15
- data/lib/rumale/preprocessing/label_binarizer.rb +0 -15
- data/lib/rumale/preprocessing/label_encoder.rb +0 -15
- data/lib/rumale/preprocessing/max_abs_scaler.rb +0 -15
- data/lib/rumale/preprocessing/min_max_scaler.rb +0 -17
- data/lib/rumale/preprocessing/one_hot_encoder.rb +0 -19
- data/lib/rumale/preprocessing/ordinal_encoder.rb +0 -13
- data/lib/rumale/preprocessing/standard_scaler.rb +0 -15
- data/lib/rumale/version.rb +1 -1
- metadata +8 -3
- data/lib/rumale/naive_bayes/naive_bayes.rb +0 -250
@@ -0,0 +1,82 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'rumale/naive_bayes/base_naive_bayes'
|
4
|
+
|
5
|
+
module Rumale
|
6
|
+
module NaiveBayes
|
7
|
+
# BernoulliNB is a class that implements Bernoulli Naive Bayes classifier.
|
8
|
+
#
|
9
|
+
# @example
|
10
|
+
# estimator = Rumale::NaiveBayes::BernoulliNB.new(smoothing_param: 1.0, bin_threshold: 0.0)
|
11
|
+
# estimator.fit(training_samples, training_labels)
|
12
|
+
# results = estimator.predict(testing_samples)
|
13
|
+
#
|
14
|
+
# *Reference*
|
15
|
+
# - C D. Manning, P. Raghavan, and H. Schutze, "Introduction to Information Retrieval," Cambridge University Press., 2008.
|
16
|
+
class BernoulliNB < BaseNaiveBayes
|
17
|
+
# Return the class labels.
|
18
|
+
# @return [Numo::Int32] (size: n_classes)
|
19
|
+
attr_reader :classes
|
20
|
+
|
21
|
+
# Return the prior probabilities of the classes.
|
22
|
+
# @return [Numo::DFloat] (shape: [n_classes])
|
23
|
+
attr_reader :class_priors
|
24
|
+
|
25
|
+
# Return the conditional probabilities for features of each class.
|
26
|
+
# @return [Numo::DFloat] (shape: [n_classes, n_features])
|
27
|
+
attr_reader :feature_probs
|
28
|
+
|
29
|
+
# Create a new classifier with Bernoulli Naive Bayes.
|
30
|
+
#
|
31
|
+
# @param smoothing_param [Float] The Laplace smoothing parameter.
|
32
|
+
# @param bin_threshold [Float] The threshold for binarizing of features.
|
33
|
+
def initialize(smoothing_param: 1.0, bin_threshold: 0.0)
|
34
|
+
check_params_numeric(smoothing_param: smoothing_param, bin_threshold: bin_threshold)
|
35
|
+
check_params_positive(smoothing_param: smoothing_param)
|
36
|
+
@params = {}
|
37
|
+
@params[:smoothing_param] = smoothing_param
|
38
|
+
@params[:bin_threshold] = bin_threshold
|
39
|
+
end
|
40
|
+
|
41
|
+
# Fit the model with given training data.
|
42
|
+
#
|
43
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The training data to be used for fitting the model.
|
44
|
+
# @param y [Numo::Int32] (shape: [n_samples]) The categorical variables (e.g. labels)
|
45
|
+
# to be used for fitting the model.
|
46
|
+
# @return [BernoulliNB] The learned classifier itself.
|
47
|
+
def fit(x, y)
|
48
|
+
x = check_convert_sample_array(x)
|
49
|
+
y = check_convert_label_array(y)
|
50
|
+
check_sample_label_size(x, y)
|
51
|
+
n_samples, = x.shape
|
52
|
+
bin_x = Numo::DFloat[*x.gt(@params[:bin_threshold])]
|
53
|
+
@classes = Numo::Int32[*y.to_a.uniq.sort]
|
54
|
+
n_samples_each_class = Numo::DFloat[*@classes.to_a.map { |l| y.eq(l).count.to_f }]
|
55
|
+
@class_priors = n_samples_each_class / n_samples
|
56
|
+
count_features = Numo::DFloat[*@classes.to_a.map { |l| bin_x[y.eq(l).where, true].sum(0) }]
|
57
|
+
count_features += @params[:smoothing_param]
|
58
|
+
n_samples_each_class += 2.0 * @params[:smoothing_param]
|
59
|
+
n_classes = @classes.size
|
60
|
+
@feature_probs = count_features / n_samples_each_class.reshape(n_classes, 1)
|
61
|
+
self
|
62
|
+
end
|
63
|
+
|
64
|
+
# Calculate confidence scores for samples.
|
65
|
+
#
|
66
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to compute the scores.
|
67
|
+
# @return [Numo::DFloat] (shape: [n_samples, n_classes]) Confidence scores per sample for each class.
|
68
|
+
def decision_function(x)
|
69
|
+
x = check_convert_sample_array(x)
|
70
|
+
n_classes = @classes.size
|
71
|
+
bin_x = Numo::DFloat[*x.gt(@params[:bin_threshold])]
|
72
|
+
not_bin_x = Numo::DFloat[*x.le(@params[:bin_threshold])]
|
73
|
+
log_likelihoods = Array.new(n_classes) do |l|
|
74
|
+
Math.log(@class_priors[l]) + (
|
75
|
+
(Numo::DFloat[*bin_x] * Numo::NMath.log(@feature_probs[l, true])).sum(1)
|
76
|
+
(Numo::DFloat[*not_bin_x] * Numo::NMath.log(1.0 - @feature_probs[l, true])).sum(1))
|
77
|
+
end
|
78
|
+
Numo::DFloat[*log_likelihoods].transpose
|
79
|
+
end
|
80
|
+
end
|
81
|
+
end
|
82
|
+
end
|
@@ -0,0 +1,85 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'rumale/naive_bayes/base_naive_bayes'
|
4
|
+
|
5
|
+
module Rumale
|
6
|
+
module NaiveBayes
|
7
|
+
# ComplementNB is a class that implements Complement Naive Bayes classifier.
|
8
|
+
#
|
9
|
+
# @example
|
10
|
+
# estimator = Rumale::NaiveBayes::ComplementNB.new(smoothing_param: 1.0)
|
11
|
+
# estimator.fit(training_samples, training_labels)
|
12
|
+
# results = estimator.predict(testing_samples)
|
13
|
+
#
|
14
|
+
# *Reference*
|
15
|
+
# - Rennie, J. D. M., Shih, L., Teevan, J., and Karger, D. R., "Tackling the Poor Assumptions of Naive Bayes Text Classifiers," ICML' 03, pp. 616--623, 2013.
|
16
|
+
class ComplementNB < BaseNaiveBayes
|
17
|
+
# Return the class labels.
|
18
|
+
# @return [Numo::Int32] (size: n_classes)
|
19
|
+
attr_reader :classes
|
20
|
+
|
21
|
+
# Return the prior probabilities of the classes.
|
22
|
+
# @return [Numo::DFloat] (shape: [n_classes])
|
23
|
+
attr_reader :class_priors
|
24
|
+
|
25
|
+
# Return the conditional probabilities for features of each class.
|
26
|
+
# @return [Numo::DFloat] (shape: [n_classes, n_features])
|
27
|
+
attr_reader :feature_probs
|
28
|
+
|
29
|
+
# Create a new classifier with Complement Naive Bayes.
|
30
|
+
#
|
31
|
+
# @param smoothing_param [Float] The smoothing parameter.
|
32
|
+
# @param norm [Boolean] The flag indicating whether to normlize the weight vectors.
|
33
|
+
def initialize(smoothing_param: 1.0, norm: false)
|
34
|
+
check_params_numeric(smoothing_param: smoothing_param)
|
35
|
+
check_params_positive(smoothing_param: smoothing_param)
|
36
|
+
check_params_boolean(norm: norm)
|
37
|
+
@params = {}
|
38
|
+
@params[:smoothing_param] = smoothing_param
|
39
|
+
@params[:norm] = norm
|
40
|
+
end
|
41
|
+
|
42
|
+
# Fit the model with given training data.
|
43
|
+
#
|
44
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The training data to be used for fitting the model.
|
45
|
+
# @param y [Numo::Int32] (shape: [n_samples]) The categorical variables (e.g. labels)
|
46
|
+
# to be used for fitting the model.
|
47
|
+
# @return [ComplementNB] The learned classifier itself.
|
48
|
+
def fit(x, y)
|
49
|
+
x = check_convert_sample_array(x)
|
50
|
+
y = check_convert_label_array(y)
|
51
|
+
check_sample_label_size(x, y)
|
52
|
+
n_samples, = x.shape
|
53
|
+
@classes = Numo::Int32[*y.to_a.uniq.sort]
|
54
|
+
@class_priors = Numo::DFloat[*@classes.to_a.map { |l| y.eq(l).count.fdiv(n_samples) }]
|
55
|
+
@class_log_probs = Numo::NMath.log(@class_priors)
|
56
|
+
compl_features = Numo::DFloat[*@classes.to_a.map { |l| x[y.ne(l).where, true].sum(0) }]
|
57
|
+
compl_features += @params[:smoothing_param]
|
58
|
+
n_classes = @classes.size
|
59
|
+
@feature_probs = compl_features / compl_features.sum(1).reshape(n_classes, 1)
|
60
|
+
feature_log_probs = Numo::NMath.log(@feature_probs)
|
61
|
+
@weights = if normalize?
|
62
|
+
feature_log_probs / feature_log_probs.sum(1).reshape(n_classes, 1)
|
63
|
+
else
|
64
|
+
-feature_log_probs
|
65
|
+
end
|
66
|
+
self
|
67
|
+
end
|
68
|
+
|
69
|
+
# Calculate confidence scores for samples.
|
70
|
+
#
|
71
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to compute the scores.
|
72
|
+
# @return [Numo::DFloat] (shape: [n_samples, n_classes]) Confidence scores per sample for each class.
|
73
|
+
def decision_function(x)
|
74
|
+
x = check_convert_sample_array(x)
|
75
|
+
@class_log_probs + x.dot(@weights.transpose)
|
76
|
+
end
|
77
|
+
|
78
|
+
private
|
79
|
+
|
80
|
+
def normalize?
|
81
|
+
@params[:norm] == true
|
82
|
+
end
|
83
|
+
end
|
84
|
+
end
|
85
|
+
end
|
@@ -0,0 +1,69 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'rumale/naive_bayes/base_naive_bayes'
|
4
|
+
|
5
|
+
module Rumale
|
6
|
+
module NaiveBayes
|
7
|
+
# GaussianNB is a class that implements Gaussian Naive Bayes classifier.
|
8
|
+
#
|
9
|
+
# @example
|
10
|
+
# estimator = Rumale::NaiveBayes::GaussianNB.new
|
11
|
+
# estimator.fit(training_samples, training_labels)
|
12
|
+
# results = estimator.predict(testing_samples)
|
13
|
+
class GaussianNB < BaseNaiveBayes
|
14
|
+
# Return the class labels.
|
15
|
+
# @return [Numo::Int32] (size: n_classes)
|
16
|
+
attr_reader :classes
|
17
|
+
|
18
|
+
# Return the prior probabilities of the classes.
|
19
|
+
# @return [Numo::DFloat] (shape: [n_classes])
|
20
|
+
attr_reader :class_priors
|
21
|
+
|
22
|
+
# Return the mean vectors of the classes.
|
23
|
+
# @return [Numo::DFloat] (shape: [n_classes, n_features])
|
24
|
+
attr_reader :means
|
25
|
+
|
26
|
+
# Return the variance vectors of the classes.
|
27
|
+
# @return [Numo::DFloat] (shape: [n_classes, n_features])
|
28
|
+
attr_reader :variances
|
29
|
+
|
30
|
+
# Create a new classifier with Gaussian Naive Bayes.
|
31
|
+
def initialize
|
32
|
+
@params = {}
|
33
|
+
end
|
34
|
+
|
35
|
+
# Fit the model with given training data.
|
36
|
+
#
|
37
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The training data to be used for fitting the model.
|
38
|
+
# @param y [Numo::Int32] (shape: [n_samples]) The categorical variables (e.g. labels)
|
39
|
+
# to be used for fitting the model.
|
40
|
+
# @return [GaussianNB] The learned classifier itself.
|
41
|
+
def fit(x, y)
|
42
|
+
x = check_convert_sample_array(x)
|
43
|
+
y = check_convert_label_array(y)
|
44
|
+
check_sample_label_size(x, y)
|
45
|
+
n_samples, = x.shape
|
46
|
+
@classes = Numo::Int32[*y.to_a.uniq.sort]
|
47
|
+
@class_priors = Numo::DFloat[*@classes.to_a.map { |l| y.eq(l).count / n_samples.to_f }]
|
48
|
+
@means = Numo::DFloat[*@classes.to_a.map { |l| x[y.eq(l).where, true].mean(0) }]
|
49
|
+
@variances = Numo::DFloat[*@classes.to_a.map { |l| x[y.eq(l).where, true].var(0) }]
|
50
|
+
self
|
51
|
+
end
|
52
|
+
|
53
|
+
# Calculate confidence scores for samples.
|
54
|
+
#
|
55
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to compute the scores.
|
56
|
+
# @return [Numo::DFloat] (shape: [n_samples, n_classes]) Confidence scores per sample for each class.
|
57
|
+
def decision_function(x)
|
58
|
+
x = check_convert_sample_array(x)
|
59
|
+
n_classes = @classes.size
|
60
|
+
log_likelihoods = Array.new(n_classes) do |l|
|
61
|
+
Math.log(@class_priors[l]) - 0.5 * (
|
62
|
+
Numo::NMath.log(2.0 * Math::PI * @variances[l, true]) +
|
63
|
+
((x - @means[l, true])**2 / @variances[l, true])).sum(1)
|
64
|
+
end
|
65
|
+
Numo::DFloat[*log_likelihoods].transpose
|
66
|
+
end
|
67
|
+
end
|
68
|
+
end
|
69
|
+
end
|
@@ -0,0 +1,74 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'rumale/naive_bayes/base_naive_bayes'
|
4
|
+
|
5
|
+
module Rumale
|
6
|
+
module NaiveBayes
|
7
|
+
# MultinomialNB is a class that implements Multinomial Naive Bayes classifier.
|
8
|
+
#
|
9
|
+
# @example
|
10
|
+
# estimator = Rumale::NaiveBayes::MultinomialNB.new(smoothing_param: 1.0)
|
11
|
+
# estimator.fit(training_samples, training_labels)
|
12
|
+
# results = estimator.predict(testing_samples)
|
13
|
+
#
|
14
|
+
# *Reference*
|
15
|
+
# - C D. Manning, P. Raghavan, and H. Schutze, "Introduction to Information Retrieval," Cambridge University Press., 2008.
|
16
|
+
class MultinomialNB < BaseNaiveBayes
|
17
|
+
# Return the class labels.
|
18
|
+
# @return [Numo::Int32] (size: n_classes)
|
19
|
+
attr_reader :classes
|
20
|
+
|
21
|
+
# Return the prior probabilities of the classes.
|
22
|
+
# @return [Numo::DFloat] (shape: [n_classes])
|
23
|
+
attr_reader :class_priors
|
24
|
+
|
25
|
+
# Return the conditional probabilities for features of each class.
|
26
|
+
# @return [Numo::DFloat] (shape: [n_classes, n_features])
|
27
|
+
attr_reader :feature_probs
|
28
|
+
|
29
|
+
# Create a new classifier with Multinomial Naive Bayes.
|
30
|
+
#
|
31
|
+
# @param smoothing_param [Float] The Laplace smoothing parameter.
|
32
|
+
def initialize(smoothing_param: 1.0)
|
33
|
+
check_params_numeric(smoothing_param: smoothing_param)
|
34
|
+
check_params_positive(smoothing_param: smoothing_param)
|
35
|
+
@params = {}
|
36
|
+
@params[:smoothing_param] = smoothing_param
|
37
|
+
end
|
38
|
+
|
39
|
+
# Fit the model with given training data.
|
40
|
+
#
|
41
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The training data to be used for fitting the model.
|
42
|
+
# @param y [Numo::Int32] (shape: [n_samples]) The categorical variables (e.g. labels)
|
43
|
+
# to be used for fitting the model.
|
44
|
+
# @return [MultinomialNB] The learned classifier itself.
|
45
|
+
def fit(x, y)
|
46
|
+
x = check_convert_sample_array(x)
|
47
|
+
y = check_convert_label_array(y)
|
48
|
+
check_sample_label_size(x, y)
|
49
|
+
n_samples, = x.shape
|
50
|
+
@classes = Numo::Int32[*y.to_a.uniq.sort]
|
51
|
+
@class_priors = Numo::DFloat[*@classes.to_a.map { |l| y.eq(l).count / n_samples.to_f }]
|
52
|
+
count_features = Numo::DFloat[*@classes.to_a.map { |l| x[y.eq(l).where, true].sum(0) }]
|
53
|
+
count_features += @params[:smoothing_param]
|
54
|
+
n_classes = @classes.size
|
55
|
+
@feature_probs = count_features / count_features.sum(1).reshape(n_classes, 1)
|
56
|
+
self
|
57
|
+
end
|
58
|
+
|
59
|
+
# Calculate confidence scores for samples.
|
60
|
+
#
|
61
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to compute the scores.
|
62
|
+
# @return [Numo::DFloat] (shape: [n_samples, n_classes]) Confidence scores per sample for each class.
|
63
|
+
def decision_function(x)
|
64
|
+
x = check_convert_sample_array(x)
|
65
|
+
n_classes = @classes.size
|
66
|
+
bin_x = x.gt(0)
|
67
|
+
log_likelihoods = Array.new(n_classes) do |l|
|
68
|
+
Math.log(@class_priors[l]) + (Numo::DFloat[*bin_x] * Numo::NMath.log(@feature_probs[l, true])).sum(1)
|
69
|
+
end
|
70
|
+
Numo::DFloat[*log_likelihoods].transpose
|
71
|
+
end
|
72
|
+
end
|
73
|
+
end
|
74
|
+
end
|
@@ -0,0 +1,71 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'rumale/naive_bayes/base_naive_bayes'
|
4
|
+
|
5
|
+
module Rumale
|
6
|
+
module NaiveBayes
|
7
|
+
# NegationNB is a class that implements Negation Naive Bayes classifier.
|
8
|
+
#
|
9
|
+
# @example
|
10
|
+
# estimator = Rumale::NaiveBayes::NegationNB.new(smoothing_param: 1.0)
|
11
|
+
# estimator.fit(training_samples, training_labels)
|
12
|
+
# results = estimator.predict(testing_samples)
|
13
|
+
#
|
14
|
+
# *Reference*
|
15
|
+
# - Komiya, K., Sato, N., Fujimoto, K., and Kotani, Y., "Negation Naive Bayes for Categorization of Product Pages on the Web," RANLP' 11, pp. 586--592, 2011.
|
16
|
+
class NegationNB < BaseNaiveBayes
|
17
|
+
# Return the class labels.
|
18
|
+
# @return [Numo::Int32] (size: n_classes)
|
19
|
+
attr_reader :classes
|
20
|
+
|
21
|
+
# Return the prior probabilities of the classes.
|
22
|
+
# @return [Numo::DFloat] (shape: [n_classes])
|
23
|
+
attr_reader :class_priors
|
24
|
+
|
25
|
+
# Return the conditional probabilities for features of each class.
|
26
|
+
# @return [Numo::DFloat] (shape: [n_classes, n_features])
|
27
|
+
attr_reader :feature_probs
|
28
|
+
|
29
|
+
# Create a new classifier with Complement Naive Bayes.
|
30
|
+
#
|
31
|
+
# @param smoothing_param [Float] The smoothing parameter.
|
32
|
+
def initialize(smoothing_param: 1.0)
|
33
|
+
check_params_numeric(smoothing_param: smoothing_param)
|
34
|
+
check_params_positive(smoothing_param: smoothing_param)
|
35
|
+
@params = {}
|
36
|
+
@params[:smoothing_param] = smoothing_param
|
37
|
+
end
|
38
|
+
|
39
|
+
# Fit the model with given training data.
|
40
|
+
#
|
41
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The training data to be used for fitting the model.
|
42
|
+
# @param y [Numo::Int32] (shape: [n_samples]) The categorical variables (e.g. labels)
|
43
|
+
# to be used for fitting the model.
|
44
|
+
# @return [ComplementNB] The learned classifier itself.
|
45
|
+
def fit(x, y)
|
46
|
+
x = check_convert_sample_array(x)
|
47
|
+
y = check_convert_label_array(y)
|
48
|
+
check_sample_label_size(x, y)
|
49
|
+
n_samples, = x.shape
|
50
|
+
@classes = Numo::Int32[*y.to_a.uniq.sort]
|
51
|
+
@class_priors = Numo::DFloat[*@classes.to_a.map { |l| y.eq(l).count.fdiv(n_samples) }]
|
52
|
+
@class_log_probs = Numo::NMath.log(1 / (1 - @class_priors))
|
53
|
+
compl_features = Numo::DFloat[*@classes.to_a.map { |l| x[y.ne(l).where, true].sum(0) }]
|
54
|
+
compl_features += @params[:smoothing_param]
|
55
|
+
n_classes = @classes.size
|
56
|
+
@feature_probs = compl_features / compl_features.sum(1).reshape(n_classes, 1)
|
57
|
+
@weights = Numo::NMath.log(@feature_probs)
|
58
|
+
self
|
59
|
+
end
|
60
|
+
|
61
|
+
# Calculate confidence scores for samples.
|
62
|
+
#
|
63
|
+
# @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to compute the scores.
|
64
|
+
# @return [Numo::DFloat] (shape: [n_samples, n_classes]) Confidence scores per sample for each class.
|
65
|
+
def decision_function(x)
|
66
|
+
x = check_convert_sample_array(x)
|
67
|
+
@class_log_probs - x.dot(@weights.transpose)
|
68
|
+
end
|
69
|
+
end
|
70
|
+
end
|
71
|
+
end
|
@@ -127,25 +127,6 @@ module Rumale
|
|
127
127
|
n_samples = x.shape[0]
|
128
128
|
Numo::Int32.asarray(Array.new(n_samples) { |n| @classes[decision_values[n, true].max_index] })
|
129
129
|
end
|
130
|
-
|
131
|
-
# Dump marshal data.
|
132
|
-
# @return [Hash] The marshal data about KNeighborsClassifier.
|
133
|
-
def marshal_dump
|
134
|
-
{ params: @params,
|
135
|
-
prototypes: @prototypes,
|
136
|
-
labels: @labels,
|
137
|
-
classes: @classes }
|
138
|
-
end
|
139
|
-
|
140
|
-
# Load marshal data.
|
141
|
-
# @return [nil]
|
142
|
-
def marshal_load(obj)
|
143
|
-
@params = obj[:params]
|
144
|
-
@prototypes = obj[:prototypes]
|
145
|
-
@labels = obj[:labels]
|
146
|
-
@classes = obj[:classes]
|
147
|
-
nil
|
148
|
-
end
|
149
130
|
end
|
150
131
|
end
|
151
132
|
end
|
@@ -101,23 +101,6 @@ module Rumale
|
|
101
101
|
end
|
102
102
|
Numo::DFloat[*predicted_values]
|
103
103
|
end
|
104
|
-
|
105
|
-
# Dump marshal data.
|
106
|
-
# @return [Hash] The marshal data about KNeighborsRegressor.
|
107
|
-
def marshal_dump
|
108
|
-
{ params: @params,
|
109
|
-
prototypes: @prototypes,
|
110
|
-
values: @values }
|
111
|
-
end
|
112
|
-
|
113
|
-
# Load marshal data.
|
114
|
-
# @return [nil]
|
115
|
-
def marshal_load(obj)
|
116
|
-
@params = obj[:params]
|
117
|
-
@prototypes = obj[:prototypes]
|
118
|
-
@values = obj[:values]
|
119
|
-
nil
|
120
|
-
end
|
121
104
|
end
|
122
105
|
end
|
123
106
|
end
|
@@ -50,25 +50,6 @@ module Rumale
|
|
50
50
|
|
51
51
|
weight - @params[:learning_rate] * nm_fst_moment / (nm_sec_moment**0.5 + 1e-8)
|
52
52
|
end
|
53
|
-
|
54
|
-
# Dump marshal data.
|
55
|
-
# @return [Hash] The marshal data.
|
56
|
-
# def marshal_dump
|
57
|
-
# { params: @params,
|
58
|
-
# fst_moment: @fst_moment,
|
59
|
-
# sec_moment: @sec_moment,
|
60
|
-
# iter: @iter }
|
61
|
-
# end
|
62
|
-
|
63
|
-
# Load marshal data.
|
64
|
-
# @return [nil]
|
65
|
-
# def marshal_load(obj)
|
66
|
-
# @params = obj[:params]
|
67
|
-
# @fst_moment = obj[:fst_moment]
|
68
|
-
# @sec_moment = obj[:sec_moment]
|
69
|
-
# @iter = obj[:iter]
|
70
|
-
# nil
|
71
|
-
# end
|
72
53
|
end
|
73
54
|
end
|
74
55
|
end
|