Class: Rumale::Ensemble::AdaBoostRegressor

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

Overview

AdaBoostRegressor is a class that implements AdaBoost for regression. This class uses decision tree for a weak learner.

Reference

  • Shrestha, D. L., and Solomatine, D. P., “Experiments with AdaBoost.RT, an Improved Boosting Scheme for Regression,” Neural Computation 18 (7), pp. 1678–1710, 2006.

Examples:

require 'rumale/ensemble/ada_boost_regressor'

estimator =
  Rumale::Ensemble::AdaBoostRegressor.new(
    n_estimators: 10, criterion: 'mse', max_depth: 3, max_leaf_nodes: 10, min_samples_leaf: 5, random_seed: 1)
estimator.fit(training_samples, traininig_values)
results = estimator.predict(testing_samples)

Instance Attribute Summary collapse

Attributes inherited from Base::Estimator

#params

Instance Method Summary collapse

Methods included from Base::Regressor

#score

Constructor Details

#initialize(n_estimators: 10, threshold: 0.2, exponent: 1.0, criterion: 'mse', max_depth: nil, max_leaf_nodes: nil, min_samples_leaf: 1, max_features: nil, random_seed: nil) ⇒ AdaBoostRegressor

Create a new regressor with random forest.

Parameters:

  • n_estimators (Integer) (defaults to: 10) —

    The numeber of decision trees for contructing AdaBoost regressor.

  • threshold (Float) (defaults to: 0.2) —

    The threshold for delimiting correct and incorrect predictions. That is constrained to [0, 1]

  • exponent (Float) (defaults to: 1.0) —

    The exponent for the weight of each weak learner.

  • criterion (String) (defaults to: 'mse') —

    The function to evalue spliting point. Supported criteria are ‘gini’ and ‘entropy’.

  • max_depth (Integer) (defaults to: nil) —

    The maximum depth of the tree. If nil is given, decision tree grows without concern for depth.

  • max_leaf_nodes (Integer) (defaults to: nil) —

    The maximum number of leaves on decision tree. If nil is given, number of leaves is not limited.

  • min_samples_leaf (Integer) (defaults to: 1) —

    The minimum number of samples at a leaf node.

  • max_features (Integer) (defaults to: nil) —

    The number of features to consider when searching optimal split point. If nil is given, split process considers all features.

  • random_seed (Integer) (defaults to: nil) —

    The seed value using to initialize the random generator. It is used to randomly determine the order of features when deciding spliting point.



60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 60

def initialize(n_estimators: 10, threshold: 0.2, exponent: 1.0,
               criterion: 'mse', max_depth: nil, max_leaf_nodes: nil, min_samples_leaf: 1,
               max_features: nil, random_seed: nil)
  super()
  @params = {
    n_estimators: n_estimators,
    threshold: threshold,
    exponent: exponent,
    criterion: criterion,
    max_depth: max_depth,
    max_leaf_nodes: max_leaf_nodes,
    min_samples_leaf: min_samples_leaf,
    max_features: max_features,
    random_seed: random_seed || srand
  }
  @rng = Random.new(@params[:random_seed])
end

Instance Attribute Details

#estimator_weights ⇒ Numo::DFloat (readonly)

Return the weight for each weak learner.

Returns:

  • (Numo::DFloat) —

    (size: n_estimates)



35
36
37
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 35

def estimator_weights
  @estimator_weights
end

#estimators ⇒ Array<DecisionTreeRegressor> (readonly)

Return the set of estimators.

Returns:

  • (Array<DecisionTreeRegressor>) 


31
32
33
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 31

def estimators
  @estimators
end

#feature_importances ⇒ Numo::DFloat (readonly)

Return the importance for each feature.

Returns:

  • (Numo::DFloat) —

    (size: n_features)



39
40
41
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 39

def feature_importances
  @feature_importances
end

#rng ⇒ Random (readonly)

Return the random generator for random selection of feature index.

Returns:

  • (Random) 


43
44
45
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 43

def rng
  @rng
end

Instance Method Details

#fit(x, y) ⇒ AdaBoostRegressor

Fit the model with given training data.

Parameters:

  • x (Numo::DFloat) —

    (shape: [n_samples, n_features]) The training data to be used for fitting the model.

  • y (Numo::DFloat) —

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

Returns:



83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 83

def fit(x, y) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
  x = ::Rumale::Validation.check_convert_sample_array(x)
  y = ::Rumale::Validation.check_convert_target_value_array(y)
  ::Rumale::Validation.check_sample_size(x, y)
  unless y.ndim == 1
    raise ArgumentError,
          'AdaBoostRegressor supports only single-target variable regression; ' \
          'the target value array is expected to be 1-D'
  end

  # Initialize some variables.
  n_samples, n_features = x.shape
  @params[:max_features] = n_features unless @params[:max_features].is_a?(Integer)
  @params[:max_features] = [[1, @params[:max_features]].max, n_features].min # rubocop:disable Style/ComparableClamp
  observation_weights = Numo::DFloat.zeros(n_samples) + 1.fdiv(n_samples)
  @estimators = []
  @estimator_weights = []
  @feature_importances = Numo::DFloat.zeros(n_features)
  sub_rng = @rng.dup
  # Construct forest.
  @params[:n_estimators].times do |_t|
    # Fit weak learner.
    ids = ::Rumale::Utils.choice_ids(n_samples, observation_weights, sub_rng)
    tree = ::Rumale::Tree::DecisionTreeRegressor.new(
      criterion: @params[:criterion], max_depth: @params[:max_depth],
      max_leaf_nodes: @params[:max_leaf_nodes], min_samples_leaf: @params[:min_samples_leaf],
      max_features: @params[:max_features], random_seed: sub_rng.rand(::Rumale::Ensemble::Value::SEED_BASE)
    )
    tree.fit(x[ids, true], y[ids])
    pred = tree.predict(x)
    # Calculate errors.
    abs_err = ((pred - y) / y).abs
    sum_target = abs_err.gt(@params[:threshold])
    break if sum_target.count.zero?

    err = observation_weights[sum_target].sum
    break if err <= 0.0

    # Calculate weight.
    beta = err**@params[:exponent]
    weight = Math.log(1.fdiv(beta))
    # Store model.
    @estimators.push(tree)
    @estimator_weights.push(weight)
    @feature_importances += weight * tree.feature_importances
    # Update observation weights.
    update = Numo::DFloat.ones(n_samples)
    update_target = abs_err.le(@params[:threshold])
    break if update_target.count.zero?

    update[update_target] = beta
    observation_weights *= update
    observation_weights = observation_weights.clip(1.0e-15, nil)
    sum_observation_weights = observation_weights.sum
    break if sum_observation_weights.zero?

    observation_weights /= sum_observation_weights
  end
  if @estimators.empty?
    warn('Failed to converge, check hyper-parameters of AdaBoostRegressor.')
    self
  end
  @estimator_weights = Numo::DFloat.asarray(@estimator_weights)
  @feature_importances /= @estimator_weights.sum
  self
end

#predict(x) ⇒ Numo::DFloat

Predict values for samples.

Parameters:

  • x (Numo::DFloat) —

    (shape: [n_samples, n_features]) The samples to predict the values.

Returns:

  • (Numo::DFloat) —

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



154
155
156
157
158
159
160
161
162
163
164
 
# File 'rumale-ensemble/lib/rumale/ensemble/ada_boost_regressor.rb', line 154

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

  n_samples, = x.shape
  predictions = Numo::DFloat.zeros(n_samples)
  @estimators.size.times do |t|
    predictions += @estimator_weights[t] * @estimators[t].predict(x)
  end
  sum_weight = @estimator_weights.sum
  predictions / sum_weight
end