class Statsample::Multiset

Multiset joins multiple dataset with the same fields and vectors but with different number of cases. This is the base class for stratified and cluster sampling estimation

Attributes

datasets[R]

Array with Statsample::Dataset

fields[R]

Name of fields

Public Class Methods

new(fields) click to toggle source

To create a multiset

::new(%w{f1 f2 f3}) # define only fields

# File lib/statsample/multiset.rb, line 12
def initialize(fields)
    @fields=fields
    @datasets={}
end

new_empty_vectors(fields,ds_names) click to toggle source

# File lib/statsample/multiset.rb, line 16
def self.new_empty_vectors(fields,ds_names) 
    ms=Multiset.new(fields)
    ds_names.each{|d|
        ms.add_dataset(d,Dataset.new(fields))
    }
    ms
end

Public Instance Methods

[](i) click to toggle source

# File lib/statsample/multiset.rb, line 85
def[](i)
  @datasets[i]
end

add_dataset(key,ds) click to toggle source

# File lib/statsample/multiset.rb, line 59
def add_dataset(key,ds)
  if(ds.fields!=@fields)
   raise ArgumentError, "Dataset(#{ds.fields.to_s})must have the same fields of the Multiset(#{@fields})"
  else
      @datasets[key]=ds
  end
end

collect_vector(field) { |k, v| ... } click to toggle source

# File lib/statsample/multiset.rb, line 74
def collect_vector(field)
  @datasets.collect {|k,v|
    yield k, v[field]
  }
end

datasets_names() click to toggle source

# File lib/statsample/multiset.rb, line 53
def datasets_names
    @datasets.keys.sort
end

each(&block) click to toggle source

# File lib/statsample/multiset.rb, line 88
def each(&block)
  @datasets.each {|k,ds|
    next if ds.cases==0
    block.call(k,ds)
  }
end

each_vector(field) { |k, v| ... } click to toggle source

# File lib/statsample/multiset.rb, line 80
def each_vector(field)
  @datasets.each {|k,v|
    yield k, v[field]
  }
end

n_datasets() click to toggle source

# File lib/statsample/multiset.rb, line 56
def n_datasets
    @datasets.size
end

sum_field(field) { |stratum_name,vector| ... } click to toggle source

# File lib/statsample/multiset.rb, line 66
def sum_field(field)
  @datasets.inject(0) {|a,da|
    stratum_name=da[0]
    vector=da[1][field]
    val=yield stratum_name,vector
    a+val
  }
end

union() { |k,ds| ... } click to toggle source

Generate a new dataset as a union of partial dataset If block given, this is applied to each dataset before union

# File lib/statsample/multiset.rb, line 25
def union(&block)
  union_field={}
  types={}
  names={}
  labels={}
  each do |k,ds|
    if block
      ds=ds.dup
      yield k,ds
    end
    @fields.each do |f|
      union_field[f]||=Array.new
      union_field[f].concat(ds[f].data)
      types[f]||=ds[f].type
      names[f]||=ds[f].name
      labels[f]||=ds[f].labels
    end
  end
  
  @fields.each do |f|
    union_field[f]=union_field[f].to_vector(types[f])
    union_field[f].name=names[f]
    union_field[f].labels=labels[f]
  end
  ds_union=union_field.to_dataset
  ds_union.fields=@fields
  ds_union
end