statsample 0.6.5 → 0.6.7

data/lib/statsample/factor/principalaxis.rb

@@ -1,59 +1,128 @@
 module Statsample
 module Factor
+  # Principal Axis Analysis for a covariance or correlation matrix. 
+  #
+  # For PCA, use Statsample::Factor::PCA
+  # 
+  # == Usage:
+  #   require 'statsample'
+  #   a=[2.5, 0.5, 2.2, 1.9, 3.1, 2.3, 2.0, 1.0, 1.5, 1.1].to_scale
+  #   b=[2.4,0.7,2.9,2.2,3.0,2.7,1.6,1.1,1.6,0.9].to_scale
+  #   ds={'a'=>a,'b'=>b}.to_dataset
+  #   cor_matrix=Statsample::Bivariate.correlation_matrix(ds)
+  #   pa=Statsample::Factor::PrincipalAxis.new(cor_matrix)
+  #   pa.iterate(1)
+  #   pa.m
+  #   => 1
+  #   pca.component_matrix
+  #   => GSL::Matrix
+  #   [  9.622e-01 
+  #      9.622e-01 ]
+  #   pca.communalities
+  #   => [0.962964636346122, 0.962964636346122]
+  #
+  # == References:
+  #
+  # * SPSS manual
+  # * Smith, L. (2002). A tutorial on Principal Component Analysis. Available on http://courses.eas.ualberta.ca/eas570/pca_tutorial.pdf 
+  #   
   class PrincipalAxis
-    MIN_CHANGE_ESTIMATE=0.0001
+    # Minimum difference between succesive iterations on sum of communalities
+    DELTA=1e-3
+    # Maximum number of iterations
+    MAX_ITERATIONS=50
     include GetText
     bindtextdomain("statsample")
-    attr_accessor :m, :name
+    # Number of factors. Set by default to the number of factors
+    # with eigen values > 1 on PCA over data
+    attr_accessor :m
     
-    attr_reader :iterations, :initial_eigenvalues
-    def initialize(matrix ,opts=Hash.new)
+    # Name of analysis
+    attr_accessor :name
+    
+    # Number of iterations required to converge
+    attr_reader :iterations
+    # Initial eigenvalues 
+    attr_reader :initial_eigenvalues
+    # Tolerance for iteratios.
+    attr_accessor :epsilon
+    # Use SMC(squared multiple correlations) as diagonal. If false, use 1
+    attr_accessor :smc
+    # Maximum number of iterations
+    attr_accessor :max_iterations
+    # Eigenvalues of factor analysis
+    attr_accessor :eigenvalues
+    
+    def initialize(matrix, opts=Hash.new)
       @matrix=matrix
       @name=""
       @m=nil
+      @initial_eigenvalues=nil
+      @initial_communalities=nil
+      @component_matrix=nil
+      @delta=DELTA
+      @smc=true
+      @max_iterations=MAX_ITERATIONS
       opts.each{|k,v|
         self.send("#{k}=",v) if self.respond_to? k
       }
+      
+      if @m.nil?
+        pca=PCA.new(::Matrix.rows(@matrix.to_a))
+        @m=pca.m
+      end
+      
       @clean=true
     end
-    def communality(m)
+    # Communality for all variables given m factors
+    def communalities(m=nil)
       if m!=@m or @clean
         iterate(m)
-        raise "Can't calculate comunnality" if @communality.nil?
+        raise "Can't calculate comunality" if @communalities.nil?
       end
-      @communality
+      @communalities
     end
-    def component_matrix(m)
+    # Component matrix for m factors
+    def component_matrix(m=nil)
       if m!=@m  or @clean
         iterate(m)
       end
       @component_matrix
     end
-
-    def iterate(m, t=25)
+    # Iterate to find the factors
+    # Parameters
+    # * m: Number of factors
+    def iterate(m=nil)
       @clean=false
+      m||=@m
       @m=m
+      t = @max_iterations
       work_matrix=@matrix.to_a
+      
       prev_com=initial_communalities
+      
       pca=PCA.new(::Matrix.rows(work_matrix))
       @initial_eigenvalues=pca.eigenvalues
+      prev_sum=prev_com.inject(0) {|ac,v| ac+v}
       @iterations=0
       t.times do |i|
         @iterations+=1
         prev_com.each_with_index{|v,it|
           work_matrix[it][it]=v
         }
-        pca=Statsample::PCA.new(::Matrix.rows(work_matrix))
-        
-        @communality=pca.communality(m)
+        pca=PCA.new(::Matrix.rows(work_matrix))
+        @communalities=pca.communalities(m)
+        @eigenvalues=pca.eigenvalues
+        com_sum=@communalities.inject(0) {|ac,v| ac+v}
         jump=true
-        @communality.each_with_index do |v2,i2|
+        
+        break if (com_sum-prev_sum).abs<@delta
+        @communalities.each_with_index do |v2,i2|
           raise "Variable #{i2} with communality > 1" if v2>1.0
-          #p (v2-prev_com[i2]).abs
-          jump=false if (v2-prev_com[i2]).abs>=MIN_CHANGE_ESTIMATE
         end
-        break if jump
-        prev_com=@communality
+        prev_sum=com_sum
+        prev_com=@communalities
+        
       end
       @component_matrix=pca.component_matrix(m)
     end
@@ -61,14 +130,20 @@ module Factor
     
     def initial_communalities
       if @initial_communalities.nil?
+        if @smc
         @initial_communalities=@matrix.column_size.times.collect {|i|
-          rxx , rxy = FactorialAnalysis.separate_matrices(@matrix,i)
+          rxx , rxy = PrincipalAxis.separate_matrices(@matrix,i)
           matrix=(rxy.t*rxx.inverse*rxy)
           matrix[0,0]
         }
+        else
+          @initial_communalities=[1.0]*@matrix.column_size
+        end
       end      
       @initial_communalities
     end
+    
+    
     # Returns two matrixes from a correlation matrix
     # with regressors correlation matrix and criteria xy
     # matrix.
@@ -91,21 +166,22 @@ module Factor
       rxx=Matrix.rows(rows)
       [rxx,rxy]
     end
-    
-    
+    def summary
+      rp=ReportBuilder.new()
+      rp.add(self)
+      rp.to_text
+    end
     def to_reportbuilder(generator)
+      iterate if @clean
       anchor=generator.add_toc_entry(_("Factor Analysis: ")+name)
       generator.add_html "<div class='pca'>"+_("Factor Analysis")+" #{@name}<a name='#{anchor}'></a>"
-      if @m.nil?
-        # Set number of factors with eigenvalues > 1
-        m=@eigenpairs.find_all {|v| v[0]>=1.0}.size
-      else
-        m=@m
-      end
+     
       generator.add_text "Number of factors: #{m}"
+      generator.add_text "Iterations: #{@iterations}"
+      
       t=ReportBuilder::Table.new(:name=>_("Communalities"), :header=>["Variable","Initial","Extraction"])
-      communality(m).each_with_index {|com,i|
-        t.add_row([i, sprintf("%0.3f", initial_communalities[i]), sprintf("%0.3f", com)])
+      communalities(m).each_with_index {|com,i|
+        t.add_row([i, sprintf("%0.4f", initial_communalities[i]), sprintf("%0.3f", com)])
       }
       generator.parse_element(t)
       

data/lib/statsample/factor/rotation.rb

@@ -16,19 +16,36 @@ module Factor
   #   p rotation.component_transformation_matrix
   # 
   class Rotation
-    MAX_PRECISION=1e-15
+    EPSILON=1e-15
+    MAX_ITERATIONS=25
+    
     attr_reader :iterations, :rotated, :component_transformation_matrix, :h2
+    # Maximum number of iterations    
+    attr_accessor :max_iterations
+    # Maximum precision    
+    attr_accessor :epsilon
+    
+    
     def initialize(matrix, opts=Hash.new)
       @matrix=matrix
       @n=@matrix.row_size # Variables, p on original
       @m=@matrix.column_size # Factors, r on original
       @component_transformation_matrix=nil
+      @max_iterations=MAX_ITERATIONS
+      @epsilon=EPSILON
       @h2=(@matrix.collect {|c| c**2} * Matrix.column_vector([1]*@m)).column(0).to_a
+      opts.each{|k,v|
+        self.send("#{k}=",v) if self.respond_to? k
+      }
+      
+      
     end
     alias_method :communalities, :h2
     alias_method :rotated_component_matrix, :rotated
     # Start iteration of 
-    def iterate(max_i=25)
+    def iterate(max_i=nil)
+      max_i||=@max_iterations
+      @max_iterations=max_i
       t=Matrix.identity(@m)
       b=@matrix.dup
       h=Matrix.diagonal(*@h2).collect {|c| Math::sqrt(c)}
@@ -61,7 +78,7 @@ module Factor
             phi=Math::atan2(num,den) / 4.0
             # puts "#{i}-#{j}: #{phi}"
             
-            if(Math::sin(phi.abs) >= MAX_PRECISION)
+            if(Math::sin(phi.abs) >= @epsilon)
               xx_rot=( Math::cos(phi)*xx)+(Math::sin(phi)*yy)
               yy_rot=((-Math::sin(phi))*xx)+(Math::cos(phi)*yy)
               

data/lib/statsample/factor.rb

@@ -3,7 +3,7 @@ require 'statsample/factor/principalaxis'
 require 'statsample/factor/rotation'
 
 module Statsample
-  # = Factor Analysis toolbox
+  # Factor Analysis toolbox.
   # * Classes for Extraction of factors: 
   #   * Statsample::Factor::PCA
   #   * Statsample::Factor::PrincipalAxis

data/lib/statsample/graph/gdchart.rb

@@ -18,9 +18,9 @@ module Statsample
 		end
 		end
 	end
-	class Vector # :nodoc:
+	class Vector 
 		# Creates a barchart using ruby-gdchart
-		def gdchart_frequencies(file, width=300, height=150, chart_type=GDChart::BAR, options={})
+		def gdchart_frequencies(file, width=300, height=150, chart_type=GDChart::BAR, options={}) # :nodoc:
 			labels,data=[],[]
 			self.frequencies.sort.each{|k,v|
 				labels.push(k.to_s)
@@ -29,17 +29,17 @@ module Statsample
 			options['ext_color']=[0xFF3399,0xFF9933,0xFFEE33,0x33FF33, 0x9966FF]
 			Statsample::Util.chart_gdchart(file,width,height,chart_type, labels,options,1,data)
 		end
-		def gdchart_histogram(bins,file, width=300, height=150, chart_type=GDChart::BAR, options={})
-            check_type :scale
-            labels=[]
-            h=histogram(bins)
-            data=[]
-            (0...bins).each{|bin|
-                data.push(h[bin])
-                range=h.get_range(bin)
-                labels.push(((range[0]+range[1]) / 2.to_f).to_s)
-            }
-            Statsample::Util.chart_gdchart(file, width, height, chart_type, labels,options, 1,data)
+		def gdchart_histogram(bins,file, width=300, height=150, chart_type=GDChart::BAR, options={}) # :nodoc:
+      check_type :scale
+      labels=[]
+      h=histogram(bins)
+      data=[]
+      (0...bins).each{|bin|
+        data.push(h[bin])
+        range=h.get_range(bin)
+        labels.push(((range[0]+range[1]) / 2.to_f).to_s)
+      }
+      Statsample::Util.chart_gdchart(file, width, height, chart_type, labels,options, 1,data)
 		end
 	end
 end

data/lib/statsample/graph/svggraph.rb

@@ -6,178 +6,177 @@ require 'SVG/Graph/Plot'
 require 'statsample/graph/svghistogram'
 
 module Statsample
-	class Vector
-		# Creates a barchart using ruby-gdchart
-		def svggraph_frequencies(file, width=600, height=300, chart_type=SVG::Graph::BarNoOp, options={})
-			labels, data1=[],[]
-			self.frequencies.sort.each{|k,v|
-				labels.push(k.to_s)
-				data1.push(v) 
-			}
-            options[:height]=height
-            options[:width]=width
-            options[:fields]=labels  
-			graph = chart_type.new(options)
-			graph.add_data(
-            :data => data1,
-            :title => "Frequencies"
-			)
-            
-			File.open(file,"w") {|f|
-              f.puts(graph.burn)
-			}
-		end
-		def svggraph_histogram(bins, options={})
-            check_type :scale
-            options={:graph_title=>"Histogram", :show_graph_title=>true,:show_normal=>true, :mean=>self.mean, :sigma=>sdp }.merge! options
-            graph = Statsample::Graph::SvgHistogram.new(options)
-            graph.histogram=histogram(bins)
-            graph
-		end
-        # Returns a Run-Sequence Plot
-        # Reference: http://www.itl.nist.gov/div898/handbook/eda/section3/runseqpl.htm
-        def svggraph_runsequence_plot(options={})
-            check_type :scale
-            options={:graph_title=>"Run-Sequence Plot", :show_graph_title=>true, :scale_x_integers => true, :add_popups=>true }.merge! options
-            vx=(1..@data.size).to_a.to_vector(:scale)
-            vy=@data.to_vector(:scale)
-            ds={'index'=>vx,'value'=>vy}.to_dataset            
-            graph = Statsample::Graph::SvgScatterplot.new(ds,options)
-            graph.set_x('index')
-            graph.parse
-            graph
-        end
-        def svggraph_boxplot(options={})
-            check_type :scale
-            options={:graph_title=>"Boxplot", :fields=>['vector'], :show_graph_title=>true}.merge! options
-            vx=@valid_data.to_a.to_vector(:scale)
-            graph = Statsample::Graph::SvgBoxplot.new(options)
-            graph.add_data(:title=>"vector", :data=>@data.to_a)
-            graph
-        end
-        
-        def svggraph_lag_plot(options={})
-            check_type :scale
-            options={:graph_title=>"Lag Plot", :show_graph_title=>true}.merge! options
-            vx=@valid_data[0...(@valid_data.size-1)].to_vector(:scale)
-            vy=@valid_data[1...@valid_data.size].to_vector(:scale)
-            ds={'x_minus_1'=>vx,'x'=>vy}.to_dataset            
-            graph = Statsample::Graph::SvgScatterplot.new(ds,options)
-            graph.set_x('x_minus_1')
-            graph.parse
-            graph
-        end
-        
-        # Returns a Normal Probability Plot
-        # Reference: http://www.itl.nist.gov/div898/handbook/eda/section3/normprpl.htm
-        def svggraph_normalprobability_plot(options={})
-            extend Statsample::Util
-            check_type :scale
-            options={:graph_title=>"Normal Probability Plot", :show_graph_title=>true}.merge! options
-            n=@valid_data.size
-            vx=(1..@valid_data.size).to_a.collect{|i|
-                Distribution::Normal.p_value(normal_order_statistic_medians(i,n))
-            }.to_vector(:scale)
-            vy=@valid_data.sort.to_vector(:scale)
-            ds={'normal_order_statistics_medians'=>vx, 'ordered_response'=>vy}.to_dataset            
-            graph = Statsample::Graph::SvgScatterplot.new(ds,options)
-            graph.set_x('normal_order_statistics_medians')
-            graph.parse
-            graph
-        end
-	end
+  class Vector
+    # Creates a barchart using ruby-gdchart
+    def svggraph_frequencies(file, width=600, height=300, chart_type=SVG::Graph::BarNoOp, options={})
+      labels, data1=[],[]
+      self.frequencies.sort.each{|k,v|
+        labels.push(k.to_s)
+        data1.push(v)
+      }
+      options[:height]=height
+      options[:width]=width
+      options[:fields]=labels
+      graph = chart_type.new(options)
+      graph.add_data(
+      :data => data1,
+      :title => "Frequencies"
+      )
+
+      File.open(file,"w") {|f|
+        f.puts(graph.burn)
+      }
+    end
+    def svggraph_histogram(bins, options={})
+      check_type :scale
+      options={:graph_title=>"Histogram", :show_graph_title=>true,:show_normal=>true, :mean=>self.mean, :sigma=>sdp }.merge! options
+      graph = Statsample::Graph::SvgHistogram.new(options)
+      graph.histogram=histogram(bins)
+      graph
+    end
+    # Returns a Run-Sequence Plot
+    # Reference: http://www.itl.nist.gov/div898/handbook/eda/section3/runseqpl.htm
+    def svggraph_runsequence_plot(options={})
+      check_type :scale
+      options={:graph_title=>"Run-Sequence Plot", :show_graph_title=>true, :scale_x_integers => true, :add_popups=>true }.merge! options
+      vx=(1..@data.size).to_a.to_vector(:scale)
+      vy=@data.to_vector(:scale)
+      ds={'index'=>vx,'value'=>vy}.to_dataset
+      graph = Statsample::Graph::SvgScatterplot.new(ds,options)
+      graph.set_x('index')
+      graph.parse
+      graph
+    end
+    def svggraph_boxplot(options={})
+      check_type :scale
+      options={:graph_title=>"Boxplot", :fields=>['vector'], :show_graph_title=>true}.merge! options
+      vx=@valid_data.to_a.to_vector(:scale)
+      graph = Statsample::Graph::SvgBoxplot.new(options)
+      graph.add_data(:title=>"vector", :data=>@data.to_a)
+      graph
+    end
+
+    def svggraph_lag_plot(options={})
+      check_type :scale
+      options={:graph_title=>"Lag Plot", :show_graph_title=>true}.merge! options
+      vx=@valid_data[0...(@valid_data.size-1)].to_vector(:scale)
+      vy=@valid_data[1...@valid_data.size].to_vector(:scale)
+      ds={'x_minus_1'=>vx,'x'=>vy}.to_dataset
+      graph = Statsample::Graph::SvgScatterplot.new(ds,options)
+      graph.set_x('x_minus_1')
+      graph.parse
+      graph
+    end
+    # Returns a Normal Probability Plot
+    # Reference: http://www.itl.nist.gov/div898/handbook/eda/section3/normprpl.htm
+    def svggraph_normalprobability_plot(options={})
+      extend Statsample::Util
+      check_type :scale
+      options={:graph_title=>"Normal Probability Plot", :show_graph_title=>true}.merge! options
+      n=@valid_data.size
+      vx=(1..@valid_data.size).to_a.collect{|i|
+        Distribution::Normal.p_value(normal_order_statistic_medians(i,n))
+      }.to_vector(:scale)
+      vy=@valid_data.sort.to_vector(:scale)
+      ds={'normal_order_statistics_medians'=>vx, 'ordered_response'=>vy}.to_dataset
+      graph = Statsample::Graph::SvgScatterplot.new(ds,options)
+      graph.set_x('normal_order_statistics_medians')
+      graph.parse
+      graph
+    end
+  end
 end
 
 # replaces all key and fill classes with similar ones, without opacity
 # this allows rendering of svg and png on rox and gqview without problems
 module SVG #:nodoc:
-	module Graph 
-		class BarNoOp < Bar # :nodoc:
-			def get_css; SVG::Graph.get_css_standard; end
-		end
-		class BarHorizontalNoOp < BarHorizontal
-			def get_css; SVG::Graph.get_css_standard; end
-		end
-		
-		class LineNoOp < Line
-			def get_css; SVG::Graph.get_css_standard; end
-			
-		end
-        class PlotNoOp < Plot
-			def get_css; SVG::Graph.get_css_standard; end            
-        end
-		class PieNoOp < Pie
-			def get_css; SVG::Graph.get_css_standard; end
-			
-		end
-		class << self
-		def get_css_standard
-        return <<EOL
-/* default fill styles for multiple datasets (probably only use a single dataset on this graph though) */
-.key1,.fill1{
-	fill: #ff0000;
-	stroke: none;
-	stroke-width: 0.5px;	
-}
-.key2,.fill2{
-	fill: #0000ff;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key3,.fill3{
-	fill: #00ff00;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key4,.fill4{
-	fill: #ffcc00;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key5,.fill5{
-	fill: #00ccff;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key6,.fill6{
-	fill: #ff00ff;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key7,.fill7{
-	fill: #00ffff;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key8,.fill8{
-	fill: #ffff00;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key9,.fill9{
-	fill: #cc6666;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key10,.fill10{
-	fill: #663399;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key11,.fill11{
-	fill: #339900;
-	stroke: none;
-	stroke-width: 1px;	
-}
-.key12,.fill12{
-	fill: #9966FF;
-	stroke: none;
-	stroke-width: 1px;	
-}
-EOL
-      	end
-	end
-	end
+  module Graph
+    class BarNoOp < Bar # :nodoc:
+      def get_css; SVG::Graph.get_css_standard; end
+    end
+    class BarHorizontalNoOp < BarHorizontal
+      def get_css; SVG::Graph.get_css_standard; end
+    end
+
+    class LineNoOp < Line
+      def get_css; SVG::Graph.get_css_standard; end
+
+    end
+    class PlotNoOp < Plot
+      def get_css; SVG::Graph.get_css_standard; end
+    end
+    class PieNoOp < Pie
+      def get_css; SVG::Graph.get_css_standard; end
+
+    end
+    class << self
+      def get_css_standard
+        return <<-EOL
+        /* default fill styles for multiple datasets (probably only use a single dataset on this graph though) */
+        .key1,.fill1{
+          fill: #ff0000;
+          stroke: none;
+          stroke-width: 0.5px;
+        }
+        .key2,.fill2{
+          fill: #0000ff;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key3,.fill3{
+          fill: #00ff00;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key4,.fill4{
+          fill: #ffcc00;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key5,.fill5{
+          fill: #00ccff;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key6,.fill6{
+          fill: #ff00ff;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key7,.fill7{
+          fill: #00ffff;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key8,.fill8{
+          fill: #ffff00;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key9,.fill9{
+          fill: #cc6666;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key10,.fill10{
+          fill: #663399;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key11,.fill11{
+          fill: #339900;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        .key12,.fill12{
+          fill: #9966FF;
+          stroke: none;
+          stroke-width: 1px;
+        }
+        EOL
+      end
+    end
+  end
 end
 
 require 'statsample/graph/svgscatterplot'