statsample 0.7.0 → 0.8.0

data/lib/statsample/graph.rb

@@ -0,0 +1,8 @@
+require 'statsample/graph/svggraph'
+require 'statsample/graph/svghistogram'
+require 'statsample/graph/svgscatterplot'
+
+module Statsample
+  module Graph
+  end
+end

data/lib/statsample/graph/svggraph.rb

@@ -3,7 +3,6 @@ require 'SVG/Graph/BarHorizontal'
 require 'SVG/Graph/Pie'
 require 'SVG/Graph/Line'
 require 'SVG/Graph/Plot'
-require 'statsample/graph/svghistogram'
 
 module Statsample
   class Vector
@@ -183,6 +182,3 @@ module SVG #:nodoc:
     end
   end
 end
-
-require 'statsample/graph/svgscatterplot'
-require 'statsample/graph/svgboxplot'

data/lib/statsample/regression/multiple/baseengine.rb

@@ -160,34 +160,31 @@ module Statsample
           rp.add(self)
           rp.to_text
         end
-        def report_building(generator)
-          anchor=generator.toc_entry(_("Multiple Regression: ")+@name)
-          generator.html "<div class='multiple-regression'>#{@name}<a name='#{anchor}'></a>"
-          c=coeffs
-          generator.text(_("Engine: %s") % self.class)
-          generator.text(_("Cases(listwise)=%d(%d)") % [@ds.cases, @ds_valid.cases])
-          generator.text("R=#{sprintf('%0.3f',r)}")
-          generator.text("R^2=#{sprintf('%0.3f',r2)}")
-          
-          generator.text(_("Equation")+"="+ sprintf('%0.3f',constant) +" + "+ @fields.collect {|k| sprintf('%0.3f%s',c[k],k)}.join(' + ') )
-          
-          t=ReportBuilder::Table.new(:name=>"ANOVA", :header=>%w{source ss df ms f s})
-          t.row([_("Regression"), sprintf("%0.3f",ssr), df_r, sprintf("%0.3f",msr), sprintf("%0.3f",f), sprintf("%0.3f", significance)])
-          t.row([_("Error"), sprintf("%0.3f",sse), df_e, sprintf("%0.3f",mse)])
-  
-          t.row([_("Total"), sprintf("%0.3f",sst), df_r+df_e])
-          generator.parse_element(t)
-          sc=standarized_coeffs
-          cse=coeffs_se
-          t=ReportBuilder::Table.new(:name=>"Beta coefficients", :header=>%w{coeff b beta se t}.collect{|field| _(field)} )
-          
-            t.row([_("Constant"), sprintf("%0.3f", constant), "-", sprintf("%0.3f", constant_se), sprintf("%0.3f", constant_t)])
-          
-          @fields.each do |f|
-            t.row([f, sprintf("%0.3f", c[f]), sprintf("%0.3f", sc[f]), sprintf("%0.3f", cse[f]), sprintf("%0.3f", c[f].quo(cse[f]))])
-          end  
-          generator.parse_element(t)
-          generator.html("</div>")
+        def report_building(b)
+          b.section(:name=>_("Multiple Regression: ")+@name) do |g|
+            c=coeffs
+            g.text(_("Engine: %s") % self.class)
+            g.text(_("Cases(listwise)=%d(%d)") % [@ds.cases, @ds_valid.cases])
+            g.text("R=#{sprintf('%0.3f',r)}")
+            g.text("R^2=#{sprintf('%0.3f',r2)}")
+            
+            g.text(_("Equation")+"="+ sprintf('%0.3f',constant) +" + "+ @fields.collect {|k| sprintf('%0.3f%s',c[k],k)}.join(' + ') )
+            
+            g.table(:name=>"ANOVA", :header=>%w{source ss df ms f s}) do |t|
+              t.row([_("Regression"), sprintf("%0.3f",ssr), df_r, sprintf("%0.3f",msr), sprintf("%0.3f",f), sprintf("%0.3f", significance)])
+              t.row([_("Error"), sprintf("%0.3f",sse), df_e, sprintf("%0.3f",mse)])
+      
+              t.row([_("Total"), sprintf("%0.3f",sst), df_r+df_e])
+            end
+            sc=standarized_coeffs
+            cse=coeffs_se
+            g.table(:name=>"Beta coefficients", :header=>%w{coeff b beta se t}.collect{|field| _(field)} ) do |t|
+              t.row([_("Constant"), sprintf("%0.3f", constant), "-", sprintf("%0.3f", constant_se), sprintf("%0.3f", constant_t)])
+              @fields.each do |f|
+                t.row([f, sprintf("%0.3f", c[f]), sprintf("%0.3f", sc[f]), sprintf("%0.3f", cse[f]), sprintf("%0.3f", c[f].quo(cse[f]))])
+              end  
+            end
+          end
         end
         
         

data/lib/statsample/regression/multiple/matrixengine.rb

@@ -194,41 +194,37 @@ class MatrixEngine < BaseEngine
     matrix.collect {|i| Math::sqrt(i) if i>0 }[0,0]
   end
   
-  def report_building(generator) # :nodoc:
-    anchor=generator.toc_entry(_("Multiple Regression: ")+@name)
-    generator.html "<div class='multiple-regression'>#{@name}<a name='#{anchor}'></a>"
-    c=coeffs
-    generator.text(_("Engine: %s") % self.class)
-    generator.text(_("Cases=%d") % [@cases])
-    generator.text("R=#{sprintf('%0.3f',r)}")
-    generator.text("R^2=#{sprintf('%0.3f',r2)}")
-    
-    generator.text(_("Equation")+"="+ sprintf('%0.3f',constant) +" + "+ @fields.collect {|k| sprintf('%0.3f%s',c[k],k)}.join(' + ') )
-    
-    t=ReportBuilder::Table.new(:name=>"ANOVA", :header=>%w{source ss df ms f s})
-    t.row([_("Regression"), sprintf("%0.3f",ssr), df_r, sprintf("%0.3f",msr), sprintf("%0.3f",f), sprintf("%0.3f", significance)])
-    t.row([_("Error"), sprintf("%0.3f",sse), df_e, sprintf("%0.3f",mse)])
-  
-    t.row([_("Total"), sprintf("%0.3f",sst), df_r+df_e])
-    generator.parse_element(t)
-    sc=standarized_coeffs
-    cse=coeffs_se
-    t=ReportBuilder::Table.new(:name=>"Beta coefficients", :header=>%w{coeff b beta se t}.collect{|field| _(field)} )
-    
-    if (constant_se.nil?)
-      t.row([_("Constant"), sprintf("%0.3f", constant),"--","?","?"])
-    else
-      t.row([_("Constant"), sprintf("%0.3f", constant), "-", sprintf("%0.3f", constant_se), sprintf("%0.3f", constant_t)])
-    end
-    
-    @fields.each do |f|
-      t.row([f, sprintf("%0.3f", c[f]), sprintf("%0.3f", sc[f]), sprintf("%0.3f", cse[f]), sprintf("%0.3f", c[f].quo(cse[f]))])
-    end  
-    generator.parse_element(t)
-    generator.html("</div>")
+  def report_building(builder) # :nodoc:
+    builder.section(:name=>_("Multiple Regression: ")+@name) do |g|
+      c=coeffs
+      g.text(_("Engine: %s") % self.class)
+      g.text(_("Cases=%d") % [@cases])
+      g.text("R=#{sprintf('%0.3f',r)}")
+      g.text("R^2=#{sprintf('%0.3f',r2)}")
+      
+      g.text(_("Equation")+"="+ sprintf('%0.3f',constant) +" + "+ @fields.collect {|k| sprintf('%0.3f%s',c[k],k)}.join(' + ') )
+      
+      g.table(:name=>"ANOVA", :header=>%w{source ss df ms f s}) do |t|
+        t.row([_("Regression"), sprintf("%0.3f",ssr), df_r, sprintf("%0.3f",msr), sprintf("%0.3f",f), sprintf("%0.3f", significance)])
+        t.row([_("Error"), sprintf("%0.3f",sse), df_e, sprintf("%0.3f",mse)])
+      
+        t.row([_("Total"), sprintf("%0.3f",sst), df_r+df_e])
+      end
+      sc=standarized_coeffs
+      cse=coeffs_se
+      g.table(:name=>"Beta coefficients", :header=>%w{coeff b beta se t}.collect{|field| _(field)} ) do |t|
+        if (constant_se.nil?)
+          t.row([_("Constant"), sprintf("%0.3f", constant),"--","?","?"])
+        else
+          t.row([_("Constant"), sprintf("%0.3f", constant), "-", sprintf("%0.3f", constant_se), sprintf("%0.3f", constant_t)])
+        end
+        
+        @fields.each do |f|
+          t.row([f, sprintf("%0.3f", c[f]), sprintf("%0.3f", sc[f]), sprintf("%0.3f", cse[f]), sprintf("%0.3f", c[f].quo(cse[f]))])
+        end  
+      end
+    end    
   end
-  
-  
 end
 end
 end

data/lib/statsample/test.rb

@@ -4,7 +4,28 @@ module Statsample
   module Test
     autoload(:UMannWhitney, 'statsample/test/umannwhitney')
     autoload(:Levene, 'statsample/test/levene')
-
+    autoload(:T, 'statsample/test/t')
+    
+    # Returns probability of getting a value lower or higher
+    # than sample, using cdf and number of tails.
+    # * For one tail left, return the cdf
+    # * For one tail right, return 1-cdf
+    # * For two tails, returns 2*right_tail(cdf.abs)
+    def p_using_cdf(cdf, tails=:both)
+      tails=:both if tails==2
+      tails=:right if tails==1 or tails==:positive
+      tails=:left if tails==:negative
+      case tails
+        when :left then cdf
+        when :right then 1-cdf
+        when :both 
+          if cdf>=0.5
+            cdf=1-cdf
+          end
+          2*cdf
+      end
+    end
+    extend self
     # Calculate chi square for two Matrix
     class << self
       def chi_square(real,expected)
@@ -22,6 +43,20 @@ module Statsample
       def u_mannwhitney(v1p,v2p)
         Statsample::Test::UMannWhitney.new(v1p,v2p)
       end
+      # Shorthand for Statsample::Test::T::OneSample.new
+      def t_one_sample(vector, opts=Hash.new)
+        Statsample::Test::T::OneSample.new(vector,opts)
+      end
+      # Shorthand for Statsample::Test::T::TwoSamplesIndependent.new
+      def t_two_samples_independent(v1,v2, opts=Hash.new)
+        Statsample::Test::T::TwoSamplesIndependent.new(v1,v2,opts)
+      end
+
+      # Shorthand for Statsample::Test::Levene.new
+      def levene(input, opts=Hash.new)
+        Statsample::Test::Levene.new(input,opts)
+      end
+      
     end
   end
 end

data/lib/statsample/test/levene.rb

@@ -27,7 +27,11 @@ module Statsample
       attr_accessor :name
       # Input could be an array of vectors or a dataset
       def initialize(input, opts=Hash.new())
-        @vectors=input
+        if input.is_a? Statsample::Dataset
+          @vectors=input.vectors.values
+        else
+          @vectors=input
+        end
         @name="Levene Test"
         opts.each{|k,v|
           self.send("#{k}=",v) if self.respond_to? k
@@ -39,15 +43,15 @@ module Statsample
         @w
       end
       
-      def report_building(generator) # :nodoc:
-        generator.text(summary)
-        
+      def report_building(g) # :nodoc:
+        g.text @name
+        g.text "F: #{"%0.4f" % f}"
+        g.text "p: #{"%0.4f" % probability}"
+
       end
       # Summary of results
       def summary
-        "#{@name}
-F: #{f}
-p: #{probability}"
+        ReportBuilder.new(:no_title=>true).add(self).to_text
       end
       
       def compute

data/lib/statsample/test/t.rb

@@ -0,0 +1,189 @@
+module Statsample
+  module Test
+    module T
+      class << self
+        include Math
+        # Test the null hypothesis that the population mean is equal to a specified value u, one uses the statistic.
+        # Is the same formula used on t-test for paired sample.
+        # * <tt>x</tt>: sample/differences mean
+        # * <tt>u</tt>: population mean
+        # * <tt>s</tt>: sample/differences standard deviation
+        # * <tt>n</tt>: sample size
+        def one_sample(x,u,s,n)
+          (x-u).quo(s.quo(Math::sqrt(n)))
+        end
+        # Test if means of two samples are different.
+        # * <tt>x1</tt>: sample 1 mean
+        # * <tt>x2</tt>: sample 2 mean
+        # * <tt>s1</tt>: sample 1 standard deviation
+        # * <tt>s2</tt>: sample 2 standard deviation
+        # * <tt>n1</tt>: sample 1 size
+        # * <tt>n2</tt>: sample 2 size
+        # * <tt>equal_variance</tt>: true if equal_variance assumed
+        #
+        def two_sample_independent(x1, x2, s1, s2, n1, n2, equal_variance = false)
+          num=x1-x2
+          if equal_variance
+            sx1x2 = sqrt(((n1-1)*s1**2 + (n2-1)*s2**2).quo(n1+n2-2))
+            den   = sx1x2*sqrt(1.quo(n1)+1.quo(n2))
+          else
+            den=sqrt((s1**2).quo(n1) + (s2**2).quo(n2))
+          end
+          num.quo(den)
+        end
+         # Degrees of freedom for equal variance
+        def df_equal_variance(n1,n2)
+          n1+n2-2
+        end
+        # Degrees of freedom for unequal variance
+        # * <tt>s1</tt>: sample 1 standard deviation
+        # * <tt>s2</tt>: sample 2 standard deviation
+        # * <tt>n1</tt>: sample 1 size
+        # * <tt>n2</tt>: sample 2 size
+        # == Reference
+        # * http://en.wikipedia.org/wiki/Welch-Satterthwaite_equation
+        def df_not_equal_variance(s1,s2,n1,n2)
+          s2_1=s1**2
+          s2_2=s2**2
+          num=(s2_1.quo(n1)+s2_2.quo(n2))**2
+          den=(s2_1.quo(n1)**2).quo(n1-1) + (s2_2.quo(n2)**2).quo(n2-1)
+          num.quo(den)
+        end        
+      end
+      # One Sample t-test
+      # == Usage
+      #   a=1000.times.map {rand(100)}.to_scale
+      #   t_1=Statsample::Test::T::OneSample.new(a, {:u=>50})
+      #   t_1.summary
+      class OneSample
+        include Math
+        include Statsample::Test
+        extend Statsample::PromiseAfter
+        # Options
+        attr_accessor :opts
+        # Name of test
+        attr_accessor :name
+        # Population mean to contrast
+        attr_accessor :u
+        # Degress of freedom
+        attr_reader :df
+        # Value of t
+        attr_reader :t
+        # Probability
+        attr_reader :probability
+        # Tails for probability (:both, :left or :right)
+        attr_accessor :tails 
+        def initialize(vector, opts=Hash.new)
+          @vector=vector
+          default={:u=>0, :name=>"One Sample T Test", :tails=>:both}
+          @opts=default.merge(opts)
+          @name=@opts[:name]
+          @u=@opts[:u]
+          @tails=@opts[:tails]
+          @df= @vector.n_valid-1
+          @t=nil
+        end        
+
+        promise_after :compute, :t, :probability
+        
+        # Set t and probability for given u
+        def compute
+          @t  = T.one_sample(@vector.mean, @u, @vector.sd, @vector.n_valid)
+          @probability = p_using_cdf(Distribution::T.cdf(@t, @df), tails)
+        end
+        # Presents summary of analysis
+        # 
+        def summary
+          ReportBuilder.new(:no_title=>true).add(self).to_text
+        end
+        def report_building(b) # :nodoc:
+          b.section(:name=>@name) {|s|
+            s.text "Sample mean: #{@vector.mean}"
+            s.text "Population mean:#{u}"
+            s.text "Tails: #{tails}"
+            s.text sprintf("t = %0.4f, p=%0.4f, d.f=%d", t, probability, df)
+
+          }
+        end
+      end
+      # Two Sample t-test.
+      #
+      # == Usage
+      #   a=1000.times.map {rand(100)}.to_scale
+      #   b=1000.times.map {rand(100)}.to_scale
+      #   t_2=Statsample::Test::T::OneSample.new(a,b)
+      #   t_2.summary
+      
+      class TwoSamplesIndependent
+        include Math
+        include Statsample::Test
+        extend Statsample::PromiseAfter
+        # Options
+        attr_accessor :opts
+        # Name of test
+        attr_accessor :name
+        # Degress of freedom (equal variance)
+        attr_reader :df_equal_variance
+        # Degress of freedom (not equal variance)
+        attr_reader :df_not_equal_variance
+        # Value of t for equal_variance
+        attr_reader :t_equal_variance
+        # Value of t for non-equal_variance
+        attr_reader :t_not_equal_variance
+        # Probability(equal variance)
+        attr_reader :probability_equal_variance
+        # Probability(unequal variance)
+        attr_reader :probability_not_equal_variance        
+        # Tails for probability (:both, :left or :right)
+        attr_accessor :tails 
+        # Create the object
+        def initialize(v1, v2, opts=Hash.new)
+          @v1=v1
+          @v2=v2
+          default={:u=>0, :name=>"Two Sample T Test", :paired_samples=>false, :tails=>:both}
+          @opts=default.merge(opts)
+          @name=@opts[:name]
+          @tails=@opts[:tails]          
+        end        
+
+        promise_after :compute, :t_equal_variance, :t_not_equal_variance, :probability_equal_variance, :probability_not_equal_variance, :df_equal_variance, :df_not_equal_variance
+        
+        # Set t and probability for given u
+        def compute
+          @t_equal_variance= T.two_sample_independent(@v1.mean, @v2.mean, @v1.sd, @v2.sd, @v1.n_valid, @v2.n_valid,true)
+          
+          @t_not_equal_variance= T.two_sample_independent(@v1.mean, @v2.mean, @v1.sd, @v2.sd, @v1.n_valid, @v2.n_valid, false)
+
+          @df_equal_variance=T.df_equal_variance(@v1.n_valid, @v2.n_valid)
+          @df_not_equal_variance=T.df_not_equal_variance(@v1.sd, @v2.sd, @v1.n_valid, @v2.n_valid)
+          
+          @probability_equal_variance = p_using_cdf(Distribution::T.cdf(@t_equal_variance, @df_equal_variance), tails)
+          
+          @probability_not_equal_variance = p_using_cdf(Distribution::T.cdf(@t_not_equal_variance, @df_not_equal_variance), tails)
+
+        end
+        # Presents summary of analysis
+        # 
+        def summary
+          ReportBuilder.new(:no_title=>true).add(self).to_text
+        end
+        def report_building(b) # :nodoc:
+          b.section(:name=>@name) {|g|
+            g.table(:name=>"Mean and standard deviation", :header=>["Variable", "m", "sd","n"]) {|t|
+              t.row([1,"%0.4f" % @v1.mean,"%0.4f" % @v1.sd,@v1.n_valid])
+              t.row([2,"%0.4f" % @v2.mean,"%0.4f" % @v2.sd, @v2.n_valid])
+            }
+            g.section(:name=>"Levene Test") {|g1|
+              g1.parse_element(Statsample::Test.levene([@v1,@v2]))
+            }
+            
+            g.table(:name=>"T statistics",:header=>["Type","t","df", "p (#{tails} tails)"]) {|t|
+              t.row(["Equal variance", "%0.4f" % t_equal_variance, df_equal_variance, "%0.4f" % probability_equal_variance])
+              t.row(["Non equal variance", "%0.4f" % t_not_equal_variance, "%0.4f" % df_not_equal_variance, "%0.4f" % probability_not_equal_variance])
+            }
+          }
+        end
+      end      
+    end
+  end
+end

data/test/test_anova.rb

@@ -1,15 +1,13 @@
-$:.unshift(File.dirname(__FILE__)+'/../lib/')
-require 'statsample'
-require 'minitest/unit'
-MiniTest::Unit.autorun
+require(File.dirname(__FILE__)+'/test_helpers.rb')
+
 class StatsampleAnovaTestCase < MiniTest::Unit::TestCase
-	def initialize(*args)
+  def initialize(*args)
     @v1=[3,3,2,3,6].to_vector(:scale)
     @v2=[7,6,5,6,7].to_vector(:scale)
     @v3=[9,8,9,7,8].to_vector(:scale)
     @anova=Statsample::Anova::OneWay.new([@v1,@v2,@v3])
-		super
-	end
+    super
+  end
   def test_basic
     assert_in_delta(72.933, @anova.sst,0.001)
     assert_in_delta(14.8,@anova.sswg,0.001)
@@ -21,7 +19,7 @@ class StatsampleAnovaTestCase < MiniTest::Unit::TestCase
     assert_in_delta(23.568,@anova.f,0.001)
     anova2=Statsample::Anova::OneWay.new([@v1,@v1,@v1,@v1,@v2])
     assert_in_delta(3.960, anova2.f,0.001)
-		assert(@anova.significance<0.01)
-		assert_in_delta(0.016, anova2.significance,0.001)
+    assert(@anova.significance<0.01)
+    assert_in_delta(0.016, anova2.significance,0.001)
   end
-end
+end