class Statsample::Regression::Multiple::GslEngine

Class for Multiple Regression Analysis Requires rbgsl and uses a listwise aproach. Slower on prediction of values than Alglib, because predict is ruby based. Better memory management on multiple (+1000) series of regression. If you need pairwise, use RubyEngine Example:

@a=[1,3,2,4,3,5,4,6,5,7].to_vector(:scale)
@b=[3,3,4,4,5,5,6,6,4,4].to_vector(:scale)
@c=[11,22,30,40,50,65,78,79,99,100].to_vector(:scale)
@y=[3,4,5,6,7,8,9,10,20,30].to_vector(:scale)
ds={'a'=>@a,'b'=>@b,'c'=>@c,'y'=>@y}.to_dataset
lr=Statsample::Regression::Multiple::GslEngine.new(ds,'y')

@a=[1,3,2,4,3,5,4,6,5,7].to_vector(:scale)
@b=[3,3,4,4,5,5,6,6,4,4].to_vector(:scale)
@c=[11,22,30,40,50,65,78,79,99,100].to_vector(:scale)
@y=[3,4,5,6,7,8,9,10,20,30].to_vector(:scale)
ds={'a'=>@a,'b'=>@b,'c'=>@c,'y'=>@y}.to_dataset
lr=Statsample::Regression::Multiple::GslEngine.new(ds,'y')

Public Class Methods

_load(data) click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 58
def self._load(data)
  h=Marshal.load(data)
  self.new(h['ds'], h['y_var'])
end

new(ds,y_var, opts=Hash.new) click to toggle source

Calls superclass method Statsample::Regression::Multiple::BaseEngine.new

# File lib/statsample/regression/multiple/gslengine.rb, line 20
def initialize(ds,y_var, opts=Hash.new)
  super
  @ds=ds.dup_only_valid
  @ds_valid=@ds
  @valid_cases=@ds_valid.cases
  @dy=@ds[@y_var]
  @ds_indep=ds.dup(ds.fields-[y_var])
  # Create a custom matrix
  columns=[]
  @fields=[]
  max_deps = GSL::Matrix.alloc(@ds.cases, @ds.fields.size)
  constant_col=@ds.fields.size-1
  for i in 0...@ds.cases
    max_deps.set(i,constant_col,1)
  end
  j=0
  @ds.fields.each{|f|
    if f!=@y_var
      @ds[f].each_index{|i1|
        max_deps.set(i1,j,@ds[f][i1])
      }
      columns.push(@ds[f].to_a)
      @fields.push(f)
      j+=1
    end
  }
  @dep_columns=columns.dup
  @lr_s=nil
  c, @cov, @chisq, @status = GSL::MultiFit.linear(max_deps, @dy.gsl)
  @constant=c[constant_col]
  @coeffs_a=c.to_a.slice(0...constant_col)
  @coeffs=assign_names(@coeffs_a)
  c=nil
end

Public Instance Methods

_dump(i) click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 55
def _dump(i)
  Marshal.dump({'ds'=>@ds,'y_var'=>@y_var})
end

build_standarized() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 99
def build_standarized
  @ds_s=@ds.standarize
  @lr_s=GslEngine.new(@ds_s,@y_var)
end

coeffs() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 63
def coeffs
  @coeffs
end

coeffs_se() click to toggle source

Standard error for coeffs

# File lib/statsample/regression/multiple/gslengine.rb, line 116
def coeffs_se
  out={}
  evcm=estimated_variance_covariance_matrix
  @ds_valid.fields.each_with_index do |f,i|

    mi=i+1
    next if f==@y_var
    out[f]=evcm[mi,mi]
  end
  out
end

constant() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 86
def constant
  @constant
end

lr_s() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 93
def lr_s
  if @lr_s.nil?
    build_standarized
  end
  @lr_s
end

matrix_resolution() click to toggle source

Coefficients using a constant Based on www.xycoon.com/ols1.htm

# File lib/statsample/regression/multiple/gslengine.rb, line 68
def matrix_resolution
  columns=@dep_columns.dup.map {|xi| xi.map{|i| i.to_f}}
  columns.unshift([1.0]*@ds.cases)
  y=Matrix.columns([@dy.data.map  {|i| i.to_f}])
  x=Matrix.columns(columns)
  xt=x.t
  matrix=((xt*x)).inverse*xt
  matrix*y
end

process_s(v) click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 103
def process_s(v)
  lr_s.process(v)
end

r() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 80
def r
  Bivariate::pearson(@dy, predicted)
end

r2() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 77
def r2
  r**2
end

sst() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 83
def sst
  @dy.ss
end

standarized_coeffs() click to toggle source

# File lib/statsample/regression/multiple/gslengine.rb, line 89
def standarized_coeffs
  l=lr_s
  l.coeffs
end

standarized_residuals() click to toggle source

???? Not equal to SPSS output

# File lib/statsample/regression/multiple/gslengine.rb, line 107
def standarized_residuals
  res=residuals
  red_sd=residuals.sds
  res.collect {|v|
    v.quo(red_sd)
  }.to_vector(:scale)
end