Class: Core::Type::Instances

Inherits:
Object
  • Object
show all
Defined in:
lib/ruby-band/core/type/instances.rb

Overview

  • Description :

This is the main class from the Weka package for data handling. It is essentially a matrix: each row is an instance of the ‘Instance’ class, while each column is an instance of the ‘Attribute’ class The class ‘Instances’ is here extended to add custom functionalities

Direct Known Subclasses

Base

Defined Under Namespace

Classes: Base

Instance Method Summary collapse

Instance Method Details

#add_instance(instance) ⇒ Object

An Instance instance object (one row) is inserted into the current Instances object

  • Args :

    • instance -> an array of values of the correct data type (:nominal,:numeric,etc…)



199
200
201
202
203
204
205
206
207
 
# File 'lib/ruby-band/core/type/instances.rb', line 199

def add_instance(instance)
  data_ref=Array.new
  instance.each_with_index do |attribute,idx|
    data_ref << insert_attribute(attribute,idx)
  end
  double_array = data_ref.to_java :double
  single_row = Instance.new(1.0, double_array)
  self.add(single_row)
end

#add_nominal_attribute(attribute, list_values) ⇒ Object

An Attribute instance object is inserted into the current Instances object

  • Args :

    • attribute_name -> A name for the new attribute

    • values -> RubyArray with nominal values

  • WARNING :

This method only creates an empty attribute field



224
225
226
227
228
229
230
 
# File 'lib/ruby-band/core/type/instances.rb', line 224

def add_nominal_attribute(attribute,list_values)
  values = FastVector.new
  list_values.each do |val|
    values.addElement(val)         
  end
  insertAttributeAt(Attribute.new(attribute, values), self.numAttributes)
end

#add_numeric_attribute(attribute_name) ⇒ Object

An Attribute instance object is inserted into the current Instances object

  • Args :

    • attribute_name -> A name for the new attribute

  • WARNING :

This method only creates an empty attribute field



214
215
216
 
# File 'lib/ruby-band/core/type/instances.rb', line 214

def add_numeric_attribute(attribute_name)
  insertAttributeAt(Attribute.new(attribute_name), self.numAttributes)
end

#att(attr_type, name, *values) ⇒ Object

This method is used for attributes definition in uninitialized Instances-derived classes



283
284
285
286
287
288
289
 
# File 'lib/ruby-band/core/type/instances.rb', line 283

def att(attr_type,name,*values)
  att = Core::Type.create_numeric_attr(name.to_java(:string)) if attr_type == :numeric
  att = Core::Type.create_nominal_attr(name.to_java(:string),values[0]) if attr_type == :nominal
  att = Core::Type.create_date_attr(name.to_java(:string),values[0]) if attr_type == :date
  att = att = Core::Type.create_string_attr(name.to_java(:string)) if attr_type == :string      
  @positions << att
end

#check_array(data) ⇒ Object

(check function): should check that the array is bidimensional and that the lengths are equal



180
181
182
 
# File 'lib/ruby-band/core/type/instances.rb', line 180

def check_array(data)
  return true # still to be done
end

#check_numeric_instance ⇒ Object

Check if this instance’s attributes are all Numeric



77
78
79
80
81
82
83
 
# File 'lib/ruby-band/core/type/instances.rb', line 77

def check_numeric_instance
  enumerateAttributes.each do |att|
    unless att.isNumeric
      raise ArgumentError, "Sorry, attribute '#{att.name}' is not numeric!"
    end
  end
end

#date(name) ⇒ Object

This method is used for Date attributes definition in uninitialized Instances-derived classes

  • Args :

    • name -> Attribute name, a String



309
310
311
 
# File 'lib/ruby-band/core/type/instances.rb', line 309

def date(name)
  att :date, name
end

#dim ⇒ Object

Return the dimensions of the dataset (for the current Instances class object)



56
57
58
 
# File 'lib/ruby-band/core/type/instances.rb', line 56

def dim
  puts "Rows number:\t#{numInstances}\nColumns number:\t #{numAttributes}"
end

#each_column ⇒ Object 



68
69
70
 
# File 'lib/ruby-band/core/type/instances.rb', line 68

def each_column
  enumerate_attributes.each {|attribute| yield(attribute)}
end

#each_column_with_index ⇒ Object 



72
73
74
 
# File 'lib/ruby-band/core/type/instances.rb', line 72

def each_column_with_index
  enumerate_attributes.each_with_index {|attribute,id| yield(attribute,id)}
end

#each_row ⇒ Object 



60
61
62
 
# File 'lib/ruby-band/core/type/instances.rb', line 60

def each_row
  enumerate_instances.each {|inst| yield(inst)}
end

#each_row_with_index ⇒ Object 



64
65
66
 
# File 'lib/ruby-band/core/type/instances.rb', line 64

def each_row_with_index
  enumerate_instances.each_with_index {|inst,id| yield(inst,id)}
end

#mean(attribute_name) ⇒ Object

Return the mean value of a single attribute (a column from the Instances object)

  • Args :

    • attribute_name -> a String, the name of the attribute



124
125
126
127
128
129
 
# File 'lib/ruby-band/core/type/instances.rb', line 124

def mean(attribute_name) 
  sum = enumerateInstances.inject(0) do |s,x|
    s+=x.value(attribute(attribute_name))
  end
  return sum/(numInstances*1.0)
end

#merge_with(instances) ⇒ Object

Merges two sets of Instances together. The resulting set will have all the attributes of the first set plus all the attributes of the second set. The number of instances in both sets must be the same.

  • Args :

    • instances -> An Instances class object



271
272
273
 
# File 'lib/ruby-band/core/type/instances.rb', line 271

def merge_with(instances)
  return Instances.mergeInstances(self,instances)
end

#n_col ⇒ Object

Return the number of columns (Attribute objects) in the dataset



51
52
53
 
# File 'lib/ruby-band/core/type/instances.rb', line 51

def n_col
  return numAttributes
end

#n_rows ⇒ Object

Return the number of rows (Instance objects) in the dataset



46
47
48
 
# File 'lib/ruby-band/core/type/instances.rb', line 46

def n_rows
  return numInstances
end

#nominal(name, values) ⇒ Object

This method is used for Nominal attributes definition in uninitialized Instances-derived classes

  • Args :

    • name -> Attribute name, a String

    • values -> An array of values for the nominal attribute



295
296
297
 
# File 'lib/ruby-band/core/type/instances.rb', line 295

def nominal(name,values)
  att :nominal, name, values
end

#numeric(name) ⇒ Object

This method is used for Numeric attributes definition in uninitialized Instances-derived classes

  • Args :

    • name -> Attribute name, a String



302
303
304
 
# File 'lib/ruby-band/core/type/instances.rb', line 302

def numeric(name)
  att :numeric, name
end

#populate_by_row(data) ⇒ Object

An entire dataset is inserted ‘by row’ into the current Instances object i.e. one Instance object is inserted at the time

  • Args :

    • data -> a bidimensional array



188
189
190
191
192
193
194
 
# File 'lib/ruby-band/core/type/instances.rb', line 188

def populate_by_row(data)
  unless check_array(data) == false
    data.each do |row|
      add_instance(row)
    end
  end
end

#return_attr_data(att) ⇒ Object

Return data for a single attribute (a column from the Instances object)

  • Args :

    • att -> a String, the name of the attribute



106
107
108
109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/ruby-band/core/type/instances.rb', line 106

def return_attr_data(att)
  attr_values = Array.new
  if attribute(att).isNumeric
    enumerateInstances.each do |i|
      attr_values << i.value(attribute(att))
    end
  else
    attr_index = attribute(att).index
    enumerateInstances.each do |inst|
      attr_values << inst.string_value(attr_index)
    end
  end
  return attr_values
end

#string(name) ⇒ Object

This method is used for String attributes definition in uninitialized Instances-derived classes

  • Args :

    • name -> Attribute name, a String



316
317
318
 
# File 'lib/ruby-band/core/type/instances.rb', line 316

def string(name)
  att :string, name
end

#summary ⇒ Object

Print to STDOUT the list of the Instances’s attributes (with the corresponding types)



233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
 
# File 'lib/ruby-band/core/type/instances.rb', line 233

def summary
  summary = Ruport::Data::Table::new
  summary.add_column 'Attributes'
  enumerateAttributes.each_with_index do |att,idx| 
    summary.add_column idx
  end
 
  att_names = ['Names']
  enumerateAttributes.each do |att| 
    att_names << "'#{att.name}'"
  end
  summary << att_names

  att_types = ['Types']
  enumerateAttributes.each do |att|
    att_types << "Numeric" if att.isNumeric 
    att_types << "Nominal" if att.isNominal
    att_types << "Date" if att.isDate
    att_types << "String" if att.isString
  end
  summary << att_types

  display = []
  display << summary

  unless enumerate_instances.nil?
    count=0
    enumerateInstances.each {|inst| count=count+1}
    display << "\nNumber of rows: #{count}"
  end
  display 
end

#to_a2d ⇒ Object

Convert an Instances object to a bidimensional Ruby array where each row corresponds to an Instance object



26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
 
# File 'lib/ruby-band/core/type/instances.rb', line 26

def to_a2d
  matrix = Array.new
  att = Array.new
  self.enumerateAttributes.each_with_index do |a,idx|
    if a.isNumeric  
      enumerate_instances.each {|s| att << s.value(s.attribute(idx))}
      matrix << att
      att = Array.new
    else
      enumerateInstances.each do |inst|
        att << inst.string_value(idx)
      end
      matrix << att
      att = Array.new
    end
  end
  return matrix.transpose
end

#to_Apache_matrix ⇒ Object

Convert the present Instances object to an Apache matrix if every Instances attribute is Numeric



87
88
89
90
91
92
 
# File 'lib/ruby-band/core/type/instances.rb', line 87

def to_Apache_matrix
  check_numeric_instance
  ruby_array = to_a
  java_double_array = Core::Utils::bidimensional_to_double(ruby_array)
  return Core::Type::Apache_matrix.new(java_double_array)
end

#to_Apache_matrix_block ⇒ Object

Convert the present Instances object to an Apache matrix (block) if every Instances attribute is Numeric



96
97
98
99
100
101
 
# File 'lib/ruby-band/core/type/instances.rb', line 96

def to_Apache_matrix_block
  check_numeric_instance
  ruby_array = to_a
  java_double_array = Core::Utils::bidimensional_to_double(ruby_array)
  return Core::Type::Apache_matrix_block.new(java_double_array)
end

#to_ARFF(out_file) ⇒ Object

Write the content of the current Instances object to a .arff file

  • Args :

    • out_file -> a String, the name of the output file



154
155
156
157
158
159
160
 
# File 'lib/ruby-band/core/type/instances.rb', line 154

def to_ARFF(out_file)
  saver = ArffSaver.new
  saver.setInstances(self)
  out_file = File.new out_file
  saver.setFile(out_file);
  saver.writeBatch();
end

#to_CSV(out_file) ⇒ Object

Write the content of the current Instances object to a .csv file

  • Args :

    • out_file -> a String, the name of the output file



143
144
145
146
147
148
149
 
# File 'lib/ruby-band/core/type/instances.rb', line 143

def to_CSV(out_file)
  saver = CSVSaver.new
  saver.setInstances(self)
  out_file = File.new out_file
  saver.setFile(out_file);
  saver.writeBatch();
end

#to_json_format ⇒ Object

Return a json String for the current Instances object The output is modeled on the ‘datatable’ Google charts APIs More details at: ‘developers.google.com/chart/interactive/docs/reference#DataTable’



334
335
336
337
338
339
 
# File 'lib/ruby-band/core/type/instances.rb', line 334

def to_json_format
  dataset_hash = Hash.new
  dataset_hash[:cols] = enumerateAttributes.collect {|attribute| attribute.name}
  dataset_hash[:rows] = enumerateInstances.collect {|instance| instance.toString} 
  return JSON.pretty_generate(dataset_hash)
end

#variance(attribute_name) ⇒ Object

Return the variance of a single attribute (a column from the Instances object)

  • Args :

    • attribute_name -> a String, the name of the attribute



134
135
136
137
138
 
# File 'lib/ruby-band/core/type/instances.rb', line 134

def variance(attribute_name) 
  enumerateAttributes.each_with_idx do |att,idx|
    return variance(idx) if att.name==attribute_name
  end
end