statsample 0.9.0 → 0.10.0

data/lib/statsample/test/f.rb

@@ -3,34 +3,27 @@ module Statsample
     # From Wikipedia:
     # An F-test is any statistical test in which the test statistic has an F-distribution under the null hypothesis. It is most often used when comparing statistical models that have been fit to a data set, in order to identify the model that best fits the population from which the data were sampled.
     class F
-      include GetText
-      bindtextdomain("statsample")
-
       include Statsample::Test
 
-      attr_reader :ss_num, :ss_den, :df_num, :df_den, :ss_total, :df_total
+      attr_reader :var_num, :var_den, :df_num, :df_den, :var_total, :df_total
       # Tails for probability (:both, :left or :right)
       attr_accessor :tails
       # Name of F analysis
       attr_accessor :name
-      # Name of numerator
-      attr_accessor :name_numerator
-      # Name of denominator
-      attr_accessor :name_denominator
 
       # Parameters:
-      # * ss_num: explained variance / between group variance
-      # * ss_den: unexplained variance / within group variance
-      # * df_num: degrees of freedom for explained variance / k-1
-      # * df_den: degrees of freedom for unexplained variance / n-k
-      def initialize(ss_num, ss_den, df_num, df_den, opts=Hash.new)
-        @ss_num=ss_num
-        @ss_den=ss_den
+      # * var_num: variance numerator
+      # * var_den: variance denominator
+      # * df_num: degrees of freedom numerator
+      # * df_den: degrees of freedom denominator
+      def initialize(var_num, var_den, df_num, df_den, opts=Hash.new)
+        @var_num=var_num
+        @var_den=var_den
         @df_num=df_num
         @df_den=df_den
-        @ss_total=ss_num+ss_den
+        @var_total=var_num+var_den
         @df_total=df_num+df_den
-        opts_default={:tails=>:right, :name_numerator=>"Numerator", :name_denominator=>"Denominator", :name=>"F Test"}
+        opts_default={:tails=>:right, :name=>"F Test"}
         @opts=opts_default.merge(opts)
         raise "Tails should be right or left, not both" if @opts[:tails]==:both
         opts_default.keys.each {|k|
@@ -41,20 +34,17 @@ module Statsample
         ReportBuilder.new(:no_title=>true).add(self).to_text
       end
       def f
-        (@ss_num.quo(@df_num)).quo(@ss_den.quo(@df_den))
+        @var_num.quo(@var_den)
+      end
+      def to_f
+        f
       end
       # probability
       def probability
         p_using_cdf(Distribution::F.cdf(f, @df_num, @df_den), tails)
       end
-      def report_building(builder)#:nodoc:
-        builder.section(:name=>@name) do |b|
-          b.table(:name=>_("%s Table") % @name, :header=>%w{source ss df f p}.map {|v| _(v)}) do |t|
-            t.row([@name_numerator, sprintf("%0.3f",@ss_num),  @df_num,  sprintf("%0.3f",f), sprintf("%0.3f", probability)])
-            t.row([@name_denominator, sprintf("%0.3f",@ss_den), @df_den, "", ""])
-            t.row([_("Total"), sprintf("%0.3f",@ss_total), @df_total,"",""])
-          end
-        end
+      def report_building(builder) #:nodoc:
+        builder.text "%s : F(%d, %d) = %0.4f , p = %0.4f" % [@name, @df_num, @df_den, f, probability]
       end
     end
   end

data/lib/statsample/test/levene.rb

@@ -19,6 +19,7 @@ module Statsample
     # Reference:
     # * NIST/SEMATECH e-Handbook of Statistical Methods. Available on http://www.itl.nist.gov/div898/handbook/eda/section3/eda35a.htm
     class Levene
+      include Statsample::Test
       # Degrees of freedom 1 (k-1)
       attr_reader :d1
       # Degrees of freedom 2 (n-k)
@@ -42,18 +43,13 @@ module Statsample
       def f
         @w
       end
-      
-      def report_building(g) # :nodoc:
-        g.text @name
-        g.text "F: #{"%0.4f" % f}"
-        g.text "p: #{"%0.4f" % probability}"
-
+      def report_building(builder) # :nodoc:
+        builder.text "%s : F(%d, %d) = %0.4f , p = %0.4f" % [@name, @d1, @d2, f, probability]
       end
       # Summary of results
       def summary
         ReportBuilder.new(:no_title=>true).add(self).to_text
       end
-      
       def compute
         n=@vectors.inject(0) {|ac,v| ac+v.n_valid}
         
@@ -86,7 +82,7 @@ module Statsample
       # Probability.
       # With H_0 = Sum(s2)=0, probability of getting a value of the test upper or equal to the obtained on the sample
       def probability
-        1-Distribution::F.cdf(f, @d1, @d2) 
+        p_using_cdf(Distribution::F.cdf(f, @d1, @d2), :right)
       end
       
     end

data/lib/statsample/test/t.rb

@@ -31,7 +31,7 @@ module Statsample
           end
           num.quo(den)
         end
-         # Degrees of freedom for equal variance
+        # Degrees of freedom for equal variance on t test
         def df_equal_variance(n1,n2)
           n1+n2-2
         end
@@ -67,7 +67,6 @@ module Statsample
       class OneSample
         include Math
         include Statsample::Test
-        include DirtyMemoize
         # Options
         attr_accessor :opts
         # Name of test
@@ -76,15 +75,9 @@ module Statsample
         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 
         
-        dirty_writer :u, :tails
-        dirty_memoize :t, :probability
         # Create a One Sample T Test
         # Options:
         # * :u = Mean to compare. Default= 0
@@ -100,14 +93,14 @@ module Statsample
           @df= @vector.n_valid-1
           @t=nil
         end        
-       
+        def t
+          T.one_sample(@vector.mean, @u, @vector.sd, @vector.n_valid)
+        end
         
-        # 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)
+        def probability
+          p_using_cdf(Distribution::T.cdf(t, @df), tails)
         end
-        # Presents summary of analysis
+        # Summary of analysis
         # 
         def summary
           ReportBuilder.new(:no_title=>true).add(self).to_text
@@ -154,6 +147,8 @@ module Statsample
         include Math
         include Statsample::Test
         include DirtyMemoize
+        include GetText
+        bindtextdomain("statsample")
         # Options
         attr_accessor :opts
         # Name of test
@@ -204,25 +199,36 @@ module Statsample
           @probability_not_equal_variance = p_using_cdf(Distribution::T.cdf(@t_not_equal_variance, @df_not_equal_variance), tails)
 
         end
+        # Cohen's d is a measure of effect size. Its defined as the difference between two means divided by a standard deviation for the data
+        def d
+          n1=@v1.n_valid
+          n2=@v2.n_valid
+          num=@v1.mean-@v2.mean
+          den=Math::sqrt( ((n1-1)*@v1.sd+(n2-1)*@v2.sd).quo(n1+n2))
+          num.quo(den)
+        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=>_("Mean and standard deviation"), :header=>["Variable", "m", "sd","n"]) {|t|
+              t.row([@v1.name,"%0.4f" % @v1.mean,"%0.4f" % @v1.sd,@v1.n_valid])
+              t.row([@v2.name,"%0.4f" % @v2.mean,"%0.4f" % @v2.sd, @v2.n_valid])
             }
+            g.parse_element(Statsample::Test.levene([@v1,@v2],:name=>_("Levene test for equality of variances")))
             
-            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])
+            g.table(:name=>_("T statistics"),:header=>["Type","t","df", "p (#{tails} tails)"].map{|v| _(v)}) {|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])
             }
+            g.table(:name=>_("Effect size")) do |t|
+              t.row ['x1-x2', "%0.4f" % (@v1.mean-@v2.mean)]
+              t.row ['d', "%0.4f" % d]
+            end
           }
         end
       end      

data/lib/statsample/test/umannwhitney.rb

@@ -107,14 +107,15 @@ module Statsample
       attr_reader :u
       # Value of compensation for ties (useful for demostration)
       attr_reader :t
+      # Name of test
+      attr_accessor :name
       #
       # Create a new U Mann-Whitney test
       # Params: Two Statsample::Vectors
       # 
-      def initialize(v1,v2)
+      def initialize(v1,v2, opts=Hash.new)
         @n1=v1.valid_data.size
         @n2=v2.valid_data.size
-
         data=(v1.valid_data+v2.valid_data).to_scale
         groups=(([0]*@n1)+([1]*@n2)).to_vector
         ds={'g'=>groups, 'data'=>data}.to_dataset
@@ -132,11 +133,17 @@ module Statsample
         @u1=r1-((@n1*(@n1+1)).quo(2))
         @u2=r2-((@n2*(@n2+1)).quo(2))
         @u=(u1<u2) ? u1 : u2
+        opts_default={:name=>"Mann-Whitney's U"}
+        @opts=opts_default.merge(opts)
+        opts_default.keys.each {|k|
+          send("#{k}=", @opts[k])
+        }
+          
       end
       # Report results.
       def summary
         out=<<-HEREDOC
-Mann-Whitney U
+@name
 Sum of ranks v1: #{@r1.to_f}
 Sum of ranks v1: #{@r2.to_f}
 U Value: #{@u.to_f}
@@ -152,7 +159,7 @@ Z: #{sprintf("%0.3f",z)} (p: #{sprintf("%0.3f",z_probability)})
       end
       # Exact probability of finding values of U lower or equal to sample on U distribution. Use with caution with m*n>100000.
       # Uses u_sampling_distribution_as62
-      def exact_probability
+      def probability_exact
         dist=UMannWhitney.u_sampling_distribution_as62(@n1,@n2)
         sum=0
         (0..@u.to_i).each {|i|
@@ -190,9 +197,9 @@ Z: #{sprintf("%0.3f",z)} (p: #{sprintf("%0.3f",z_probability)})
         (@u-mu).quo(ou)
       end
       # Assuming H_0, the proportion of cdf with values of U lower
-      # than the sample. 
+      # than the sample, using normal approximation.
       # Use with more than 30 cases per group.
-      def z_probability
+      def probability_z
         (1-Distribution::Normal.cdf(z.abs()))*2
       end
     end

data/lib/statsample/vector.rb

@@ -7,7 +7,7 @@ class Array
 	end
     # Creates a new Statsample::Vector object of type :scale
     def to_scale(*args)
-        Statsample::Vector.new(self,:scale,*args)
+        Statsample::Vector.new(self, :scale,*args)
     end
 end
 
@@ -24,12 +24,7 @@ module Statsample
   class Vector
     include Enumerable
     include Writable
-    # DEFAULT OPTIONS
-    DEFAULT_OPTIONS={
-      :missing_values=>[],
-      :today_values=>['NOW','TODAY', :NOW, :TODAY],
-      :labels=>{}
-    }
+    include Summarizable
     # Level of measurement. Could be :nominal, :ordinal or :scale
     attr_reader :type
     # Original data. 
@@ -50,23 +45,39 @@ module Statsample
     attr_reader :gsl
     # Change label for specific values
     attr_accessor :labels
+    # Name of vector. Should be used for output by many classes
+    attr_accessor :name
+    
     #
     # Creates a new Vector object.
-    # [data]            Array of data.
-    # [type]            Level of meausurement. See Vector#type
-    # [opts]            Options
-    #   [:missing_values]  Array of missing values. See Vector#missing_values
-    #   [:today_values] Array of 'today' values. See Vector#today_values
-    #   [:labels]          Labels for data values
+    # * <tt>data</tt> Array of data.
+    # * <tt>type</tt> Level of meausurement. See Vector#type
+    # * <tt>opts</tt> Hash of options
+    #   * <tt>:missing_values</tt>  Array of missing values. See Vector#missing_values
+    #   * <tt>:today_values</tt> Array of 'today' values. See Vector#today_values
+    #   * <tt>:labels</tt> Labels for data values
+    #   * <tt>:name</tt> Name of vector
     #
     def initialize(data=[], type=:nominal, opts=Hash.new)
       raise "Data should be an array" unless data.is_a? Array
       @data=data
       @type=type
-      opts=DEFAULT_OPTIONS.merge(opts)
-      @missing_values=opts[:missing_values]
-      @labels=opts[:labels]
-      @today_values=opts[:today_values]
+      opts_default={
+        :missing_values=>[],
+        :today_values=>['NOW','TODAY', :NOW, :TODAY],
+        :labels=>{},
+        :name=>nil
+      }
+      @opts=opts_default.merge(opts)
+      if  @opts[:name].nil?
+        @@n_table||=0
+        @@n_table+=1
+        @opts[:name]="Vector #{@@n_table}"
+      end
+      @missing_values=@opts[:missing_values]
+      @labels=@opts[:labels]
+      @today_values=@opts[:today_values]
+      @name=@opts[:name]
       @valid_data=[]
       @data_with_nils=[]
       @date_data_with_nils=[]
@@ -80,12 +91,12 @@ module Statsample
     # Note: data, missing_values and labels are duplicated, so
     # changes on original vector doesn't propages to copies.
     def dup
-    Vector.new(@data.dup,@type, :missing_values => @missing_values.dup, :labels => @labels.dup)
+      Vector.new(@data.dup,@type, :missing_values => @missing_values.dup, :labels => @labels.dup, :name=>@name.dup)
     end
     # Returns an empty duplicate of the vector. Maintains the type,
     # missing values and labels.
     def dup_empty
-    Vector.new([],@type, :missing_values => @missing_values.dup, :labels => @labels.dup)
+      Vector.new([],@type, :missing_values => @missing_values.dup, :labels => @labels.dup, :name=>@name.dup)
     end
     # Raises an exception if type of vector is inferior to t type
     def check_type(t)
@@ -128,8 +139,8 @@ module Statsample
     # Vector equality.
     # Two vector will be the same if their data, missing values, type, labels are equals
     def ==(v2)
-    raise TypeError,"Argument should be a Vector" unless v2.instance_of? Statsample::Vector
-    @data==v2.data and @missing_values==v2.missing_values and @type==v2.type and @labels=v2.labels
+      raise TypeError,"Argument should be a Vector" unless v2.instance_of? Statsample::Vector
+      @data==v2.data and @missing_values==v2.missing_values and @type==v2.type and @labels==v2.labels
     end
     
     def _dump(i) # :nodoc:
@@ -189,8 +200,8 @@ module Statsample
     # Vector.set_valid_data at the end of your insertion cycle
     #
     def add(v,update_valid=true)
-    @data.push(v)
-    set_valid_data if update_valid
+      @data.push(v)
+      set_valid_data if update_valid
     end
     # Update valid_data, missing_data, data_with_nils and gsl
     # at the end of an insertion.
@@ -208,14 +219,14 @@ module Statsample
     #   v.valid_data
     #   => [2,3]
     def set_valid_data
-    @valid_data.clear
-    @missing_data.clear
-    @data_with_nils.clear
-    @date_data_with_nils.clear
-    @gsl=nil
-    set_valid_data_intern
-    set_scale_data if(@type==:scale)
-    set_date_data if(@type==:date)
+      @valid_data.clear
+      @missing_data.clear
+      @data_with_nils.clear
+      @date_data_with_nils.clear
+      @gsl=nil
+      set_valid_data_intern
+      set_scale_data if(@type==:scale)
+      set_date_data if(@type==:date)
     end
     
     if Statsample::STATSAMPLE__.respond_to?(:set_valid_data_intern)
@@ -228,21 +239,21 @@ module Statsample
     end
     end
     def _set_valid_data_intern #:nodoc:
-    @data.each do |n|
-      if is_valid? n
-        @valid_data.push(n)
-        @data_with_nils.push(n)
-      else
-        @data_with_nils.push(nil)
-        @missing_data.push(n)
+      @data.each do |n|
+        if is_valid? n
+          @valid_data.push(n)
+          @data_with_nils.push(n)
+        else
+          @data_with_nils.push(nil)
+          @missing_data.push(n)
+        end
       end
-    end
-    @has_missing_data=@missing_data.size>0
+      @has_missing_data=@missing_data.size>0
     end
     
     # Retrieves true if data has one o more missing values
     def has_missing_data?
-    @has_missing_data
+      @has_missing_data
     end
     # Retrieves label for value x. Retrieves x if
     # no label defined.
@@ -251,14 +262,14 @@ module Statsample
     end
     # Returns a Vector with data with labels replaced by the label.
     def vector_labeled
-    d=@data.collect{|x|
-      if @labels.has_key? x
-        @labels[x]
-      else
-        x
-      end
-    }
-    Vector.new(d,@type)
+      d=@data.collect{|x|
+        if @labels.has_key? x
+          @labels[x]
+        else
+          x
+        end
+      }
+      Vector.new(d,@type)
     end
     # Size of total data
     def size
@@ -427,13 +438,13 @@ module Statsample
     # In all the trails, every item have the same probability
     # of been selected.
     def sample_with_replacement(sample=1)
-    if(@type!=:scale or !Statsample.has_gsl?)
-      vds=@valid_data.size
-      (0...sample).collect{ @valid_data[rand(vds)] }
-    else
-      r = GSL::Rng.alloc(GSL::Rng::MT19937,rand(10000))
-      r.sample(@gsl, sample).to_a
-    end
+      if(@type!=:scale or !Statsample.has_gsl?)
+        vds=@valid_data.size
+        (0...sample).collect{ @valid_data[rand(vds)] }
+      else
+        r = GSL::Rng.alloc(GSL::Rng::MT19937,rand(10000))
+        r.sample(@gsl, sample).to_a
+      end
     end
     # Returns an random sample of size n, without replacement,
     # only with valid data.
@@ -597,24 +608,24 @@ module Statsample
     def proportion(v=1)
         frequencies[v].quo(@valid_data.size)
     end
-    def summary(out="")
-      out << sprintf("n valid:%d\n",n_valid)
-      out <<  sprintf("factors:%s\n",factors.join(","))
-      out <<  "mode:"+mode.to_s+"\n"
-      out <<  "Distribution:\n"
-      frequencies.sort.each{|k,v|
-        key=labels.has_key?(k) ? labels[k]:k
-        out <<  sprintf("%s : %s (%0.2f%%)\n",key,v, (v.quo(n_valid))*100)
-      }
-      if(@type==:ordinal)
-        out << "median:"+median.to_s+"\n"
-      end
-      if(@type==:scale)
-        out << "mean:"+mean.to_s+"\n"
-        out << "sd:"+sd.to_s+"\n"
-        
+    def report_building(b)
+      b.section(:name=>name) do |s|
+        s.text _("n :%d") % n        
+        s.text _("n valid:%d") % n_valid
+        s.text  _("factors:%s") % factors.join(",")
+        s.text   _("mode: %s") % mode
+        s.table(:name=>_("Distribution")) do |t|
+          frequencies.sort.each do |k,v|
+            key=labels.has_key?(k) ? labels[k]:k
+            t.row [key,v, ("%0.2f%%" % (v.quo(n_valid)*100))]
+          end
+        end
+        s.text _("median: %s") % median.to_s if(@type==:ordinal)
+        if(@type==:scale)
+          s.text _("mean: %0.4f") % mean
+          s.text _("sd: %0.4f") % sd.to_s
+        end
       end
-      out
     end
       
       # Variance of p, according to poblation size
@@ -817,8 +828,7 @@ module Statsample
           @gsl.mean
       end				
       def variance_sample(m=nil) # :nodoc:
-      check_type :scale
-          
+          check_type :scale
           m||=mean
           @gsl.variance_m
       end
@@ -881,7 +891,7 @@ module Statsample
     alias_method :sdp, :standard_deviation_population
     alias_method :sds, :standard_deviation_sample
     alias_method :cov, :coefficient_of_variation
-    alias_method :variance, :variance_sample
+    alias_method :variance, :variance_sample    
     alias_method :sd, :standard_deviation_sample
     alias_method :ss, :sum_of_squares
   end