Class: SVMKit::Preprocessing::StandardScaler

Inherits:
Object
  • Object
show all
Includes:
Base::BaseEstimator, Base::Transformer
Defined in:
lib/svmkit/preprocessing/standard_scaler.rb

Overview

Normalize samples by centering and scaling to unit variance.

normalizer = SVMKit::Preprocessing::StandardScaler.new
new_training_samples = normalizer.fit_transform(training_samples)
new_testing_samples = normalizer.transform(testing_samples)

Instance Attribute Summary collapse

Attributes included from Base::BaseEstimator

#params

Instance Method Summary collapse

Constructor Details

#initialize(_params = {}) ⇒ StandardScaler

Create a new normalizer for centering and scaling to unit variance.

:call-seq:

new() -> StandardScaler



26
27
28
29
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 26

def initialize(_params = {})
  @mean_vec = nil
  @std_vec = nil
end

Instance Attribute Details

#mean_vec ⇒ Object (readonly)

The vector consists of the mean value for each feature.



17
18
19
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 17

def mean_vec
  @mean_vec
end

#std_vec ⇒ Object (readonly)

The vector consists of the standard deviation for each feature.



20
21
22
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 20

def std_vec
  @std_vec
end

Instance Method Details

#fit(x, _y = nil) ⇒ Object

Calculate the mean value and standard deviation of each feature for scaling.

:call-seq:

fit(x) -> StandardScaler

  • Arguments :

    • x (NMatrix, shape: [n_samples, n_features]) – The samples to calculate the mean values and standard deviations.

  • Returns :

    • StandardScaler



40
41
42
43
44
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 40

def fit(x, _y = nil)
  @mean_vec = x.mean(0)
  @std_vec = x.std(0)
  self
end

#fit_transform(x, _y = nil) ⇒ Object

Calculate the mean values and standard deviations, and then normalize samples using them.

:call-seq:

fit_transform(x) -> NMatrix

  • Arguments :

    • x (NMatrix, shape: [n_samples, n_features]) – The samples to calculate the mean values and standard deviations.

  • Returns :

    • The scaled samples (NMatrix)



55
56
57
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 55

def fit_transform(x, _y = nil)
  fit(x).transform(x)
end

#marshal_dump ⇒ Object

Serializes object through Marshal#dump.



74
75
76
77
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 74

def marshal_dump # :nodoc:
  { mean_vec: Utils.dump_nmatrix(@mean_vec),
    std_vec: Utils.dump_nmatrix(@std_vec) }
end

#marshal_load(obj) ⇒ Object

Deserialize object through Marshal#load.



80
81
82
83
84
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 80

def marshal_load(obj) # :nodoc:
  @mean_vec = Utils.restore_nmatrix(obj[:mean_vec])
  @std_vec = Utils.restore_nmatrix(obj[:std_vec])
  nil
end

#transform(x) ⇒ Object

Perform standardization the given samples.

call-seq:

transform(x) -> NMatrix

  • Arguments :

    • x (NMatrix, shape: [n_samples, n_features]) – The samples to be scaled.

  • Returns :

    • The scaled samples (NMatrix)



68
69
70
71
 
# File 'lib/svmkit/preprocessing/standard_scaler.rb', line 68

def transform(x)
  n_samples, = x.shape
  (x - @mean_vec.repeat(n_samples, 0)) / @std_vec.repeat(n_samples, 0)
end