distribution 0.4.0 → 0.5.0

data/History.txt

@@ -1,3 +1,11 @@
+=== 0.5.0 / 2011-05-03
+
+* Exception raises on calculation of T's cdf with ruby engine. For now, stick to gsl implementation
+* Added logistic distribution. 
+* Raises an error on binomial if k>n. 
+* New rng for exponential distribution, based on F^-1 [clbustos]
+* Added power benchmark
+
 === 0.4.0 / 2011-02-01
 
 * Poisson and exponential distributions implemented. Implementation of inverse cdf for poisson is not perfect, yet.

data/Manifest.txt

@@ -10,6 +10,7 @@ benchmark/binomial_coefficient/experiment.rb
 benchmark/factorial_hash.rb
 benchmark/factorial_method.rb
 benchmark/odd.rb
+benchmark/power.rb
 bin/distribution
 data/template/distribution.erb
 data/template/distribution/gsl.erb
@@ -42,6 +43,8 @@ lib/distribution/hypergeometric.rb
 lib/distribution/hypergeometric/gsl.rb
 lib/distribution/hypergeometric/java.rb
 lib/distribution/hypergeometric/ruby.rb
+lib/distribution/logistic.rb
+lib/distribution/logistic/ruby.rb
 lib/distribution/math_extension.rb
 lib/distribution/normal.rb
 lib/distribution/normal/gsl.rb
@@ -64,6 +67,7 @@ spec/distribution_spec.rb
 spec/exponential_spec.rb
 spec/f_spec.rb
 spec/hypergeometric_spec.rb
+spec/logistic_spec.rb
 spec/math_extension_spec.rb
 spec/normal_spec.rb
 spec/poisson_spec.rb

data/README.txt

@@ -15,6 +15,7 @@ Includes code from statistics2 on Normal, T, F and Chi Square ruby code [http://
 * Very fast ruby 1.8.7/1.9.+ implementation, with improved method to calculate factorials and others common functions
 * All methods tested on several ranges. See spec/
 * On Jruby and Rubinius, BivariateNormal returns incorrect pdf
+* Ruby t distribution's cdf can't calculate fractional df. Install gsl, for now.
 
 == API structure
 

data/benchmark/power.rb

@@ -0,0 +1,27 @@
+require 'bench_press'
+require 'bigdecimal'
+extend BenchPress
+
+name 'Float vs Rational power'
+author 'Claudio Bustos'
+date '2011-02-02'
+summary "
+On ruby, the maximum size of a float is #{Float::MAX}. 
+With Rational, we can raise to integer numbers and surpass Float maximum.
+What is the speed reduction using Rational?"
+
+reps 1000 #number of repetitions
+int=10
+rat=10.quo(1)
+bd=BigDecimal("10")
+measure "Using float pow" do
+	int**307
+end
+
+measure "Using rational" do
+	rat**307
+end
+
+measure "Using big decimal pow" do
+	bd**307
+end

data/lib/distribution.rb

@@ -19,7 +19,7 @@
 #
 # == Other Sources
 # 
-# * Code of Ruby engines came from statistics2.rb, 
+# * Code of several Ruby engines came from statistics2.rb, 
 #   created by Shin-ichiro HARA(sinara@blade.nagaokaut.ac.jp).
 #   Retrieve from http://blade.nagaokaut.ac.jp/~sinara/ruby/math/statistics2/
 # 
@@ -49,7 +49,7 @@ require 'distribution/math_extension'
 #    Distribution::Normal.p_value(0.95)
 #    => 1.64485364660836
 module Distribution
-  VERSION="0.4.0"
+  VERSION="0.5.0"
   
   module Shorthand
     EQUIVALENCES={:p_value=>:p, :cdf=>:cdf, :pdf=>:pdf, :rng=>:r, :exact_pdf=>:epdf, :exact_cdf=>:ecdf, :exact_p_value=>:ep}
@@ -147,7 +147,7 @@ module Distribution
   autoload(:Hypergeometric, 'distribution/hypergeometric')
   autoload(:Exponential, 'distribution/exponential')
   autoload(:Poisson, 'distribution/poisson')
-
+  autoload(:Logistic, 'distribution/logistic')
 end
 
 

data/lib/distribution/binomial/ruby.rb

@@ -3,7 +3,8 @@ module Distribution
     module Ruby_
       class << self
         def pdf(k,n,pr)
-          Math.binomial_coefficient(n,k)*(pr**k)*(1-pr)**(n-k)          
+          raise "k>n" if k>n
+	  Math.binomial_coefficient(n,k)*(pr**k)*(1-pr)**(n-k)          
         end
         def cdf(k,n,pr)
           #(0..x.floor).inject(0) {|ac,i| ac+pdf(i,n,pr)}
@@ -24,4 +25,4 @@ module Distribution
       end
     end
   end
-end
+end

data/lib/distribution/exponential.rb

@@ -16,6 +16,10 @@ module Distribution
     SHORTHAND='expo'
     extend Distributable
     create_distribution_methods
+    ##
+    # :singleton-method: rng(l)
+    # Returns a lambda, which generates exponential variates
+    # with +l+ as rate parameter
 
     ##
     # :singleton-method: pdf(x,l)

data/lib/distribution/exponential/ruby.rb

@@ -2,7 +2,10 @@ module Distribution
   module Exponential
     module Ruby_
       class << self
-        def pdf(x,l)
+        def rng(l)
+	  lambda {p_value(rand(),l)}
+	end
+	def pdf(x,l)
           return 0 if x<0
           l*Math.exp(-l*x)
         end
@@ -16,4 +19,4 @@ module Distribution
       end
     end
   end
-end
+end

data/lib/distribution/logistic.rb

@@ -0,0 +1,40 @@
+require 'distribution/logistic/ruby'
+#require 'distribution/logistic/gsl'
+#require 'distribution/logistic/java'
+
+
+module Distribution
+  # From Wikipedia:
+  # In probability theory and statistics, the logistic distribution is a continuous probability distribution. Its cumulative distribution function is the logistic function, which appears in logistic regression and feedforward neural networks. It resembles the normal distribution in shape but has heavier tails (higher kurtosis).
+  module Logistic
+    SHORTHAND='logis'
+    extend Distributable
+    create_distribution_methods
+    ##
+    # :singleton-method: rng(u,s)
+    # Returns a proc, which returns a different logistic
+    # variate each time is called
+    # * u: mean
+    # * s: variance related parameter
+    ##
+    # :singleton-method: pdf(x , u,s)
+    # Returns the pdf for logistic distribution (f(x,u,s))
+    # * u: mean
+    # * s: variance related parameter
+    
+    ##
+    # :singleton-method: cdf(x , u,s)
+    # Returns the cdf for logistic distribution (F(x,u,s))
+    # * u: mean
+    # * s: variance related parameter
+
+    ##
+    # :singleton-method: p_value(pr , u,s)
+    # Returns the inverse cdf for logistic distribution 
+    # (F^-1(pr,u,s))
+    # * u: mean
+    # * s: variance related parameter
+    
+
+  end
+end

data/lib/distribution/logistic/ruby.rb

@@ -0,0 +1,20 @@
+module Distribution
+  module Logistic
+    module Ruby_
+      class << self
+	def rng(u,s)
+	  lambda {p_value(rand(),u,s)}
+	end
+        def pdf(x,u,s )
+          (Math.exp(-(x-u)/s))/(s*(1+Math.exp(-(x-u)/s)**2))
+        end
+        def cdf(x,u,s )
+          1/(1+Math.exp(-(x-u)/s))
+        end
+        def p_value(pr,u,s )
+          u+s*Math.log(pr/(1-pr))
+        end
+      end
+    end
+  end
+end

data/lib/distribution/math_extension.rb

@@ -203,6 +203,9 @@ module Distribution
     def beta(x,y)
       (gamma(x)*gamma(y)).quo(gamma(x+y))
     end
+    def logbeta(x,y)
+      (loggamma(x)+loggamma(y))-loggamma(x+y)
+    end
     # I_x(a,b): Regularized incomplete beta function
     # TODO: Find a faster version.
     # Source:
@@ -303,13 +306,13 @@ end
 
 module Math
   include Distribution::MathExtension
-  module_function :factorial, :beta, :loggamma, :binomial_coefficient, :binomial_coefficient_gamma, :regularized_beta_function, :incomplete_beta, :permutations, :rising_factorial , :fast_factorial, :combinations
+  module_function :factorial, :beta, :loggamma, :binomial_coefficient, :binomial_coefficient_gamma, :regularized_beta_function, :incomplete_beta, :permutations, :rising_factorial , :fast_factorial, :combinations, :logbeta
 end
 
 # Necessary on Ruby 1.9
 module CMath # :nodoc:
   include Distribution::MathExtension
-  module_function :factorial, :beta, :loggamma,  :binomial_coefficient, :binomial_coefficient_gamma, :regularized_beta_function, :incomplete_beta, :permutations, :rising_factorial, :fast_factorial, :combinations
+  module_function :factorial, :beta, :loggamma,  :binomial_coefficient, :binomial_coefficient_gamma, :regularized_beta_function, :incomplete_beta, :permutations, :rising_factorial, :fast_factorial, :combinations, :logbeta
 end
 
 if RUBY_VERSION<"1.9"

data/lib/distribution/t/ruby.rb

@@ -14,6 +14,10 @@ module Distribution
       # t-distribution ([1])
       # (-\infty, x]
       def p_t(df, t)
+        if df.to_i!=df
+          raise "Can't calculate not integer df"
+        end
+        df=df.to_i
         c2 = df.to_f / (df + t * t);
         s = Math.sqrt(1.0 - c2)
         s = -s if t < 0.0
@@ -24,6 +28,7 @@ module Distribution
           s *= (i - 1) * c2 / i
           i += 2
         end
+        
         if df.is_a? Float or df & 1 != 0
           0.5+(p*Math.sqrt(c2)+Math.atan(t/Math.sqrt(df))) / Math::PI
         else

data/spec/logistic_spec.rb

@@ -0,0 +1,68 @@
+require File.expand_path(File.dirname(__FILE__)+"/spec_helper.rb")
+  
+describe Distribution::Logistic do
+    
+  shared_examples_for "logistic engine" do
+    it "should return correct pdf" do
+      if @engine.respond_to? :pdf
+        1.upto(10) {
+	u=rand()
+        s=rand()+1
+        x=rand()*100-50
+	exp=Math.exp(-(x-u)/s)/(s*(1+Math.exp(-(x-u)/s)**2))
+	@engine.pdf(x,u,s).should eq(exp)
+        }
+      else
+        pending("No #{@engine}.pdf")
+      end
+    end
+    
+    it "should return correct cdf" do
+      if @engine.respond_to? :cdf
+
+
+        1.upto(100) {
+	u=rand()
+        s=rand()+1
+        x=rand()*100-50
+	exp=1/(1+Math.exp(-(x-u)/s))
+	@engine.cdf(x,u,s).should eq(exp)
+        }
+      
+      else
+        pending("No #{@engine}.cdf")
+      end
+    end
+  
+    
+    it "should return correct p_value" do
+      if @engine.respond_to? :p_value
+       1.upto(9) {|i|
+       u=rand()
+       s=rand()+1
+       x=@engine.p_value(i/10,u,s)
+       @engine.cdf(x,u,s).should eq i/10
+        }
+      else
+        pending("No #{@engine}.cdf")
+      end
+    end
+  end
+  
+
+  describe "singleton" do
+    before do
+      @engine=Distribution::Logistic
+    end
+    it_should_behave_like "logistic engine"
+  end
+  
+  describe Distribution::Logistic::Ruby_ do
+    before do
+      @engine=Distribution::Logistic::Ruby_
+    end
+    it_should_behave_like "logistic engine"
+    
+  end
+    
+end

data/spec/t_spec.rb

@@ -15,6 +15,7 @@ shared_examples_for "T engine(with pdf)" do
       [-2,0.1,0.5,1,2].each{|t|
         [2,5,10].each{|n|
           @engine.pdf(t,n).should be_within(1e-6).of(GSL::Ran.tdist_pdf(t,n))
+          @engine.pdf(t,n.to_f).should be_within(1e-6).of(@engine.pdf(t,n))
 
         }
       }
@@ -28,9 +29,16 @@ shared_examples_for "T engine" do
   
   it_only_with_gsl "should return correct cdf" do
     if @engine.respond_to? :cdf
+      # Testing with R values
+      @engine.cdf(1,2).should be_within(1e-7).of(0.7886751)
+      @engine.cdf(1,2.0).should be_within(1e-7).of(0.7886751)
+      @engine.cdf(1,3.0).should be_within(1e-7).of(0.8044989)
+      
       [-2,0.1,0.5,1,2].each{|t|
         [2,5,10].each{|n|
           @engine.cdf(t,n).should be_within(1e-4).of(GSL::Cdf.tdist_P(t,n))
+          @engine.cdf(t,n).should be_within(1e-4).of(@engine.cdf(t,n.to_f))
+          
         }
       }
     else
@@ -65,17 +73,27 @@ end
       @engine=Distribution::T::Ruby_
     end
     it_should_behave_like "T engine"    
-    it_should_behave_like "T engine(with pdf)"    
+    it_should_behave_like "T engine(with pdf)"
+    it "raise error with fractional df on cdf" do
+      lambda {@engine.cdf(1,2.5)}.should raise_error
+    end
+      it "return correct values for cdf with fractional df" do
+        pending("we have to implement partial beta");
+      end   
   end
   if Distribution.has_gsl?
     describe Distribution::T::GSL_ do
       before do
         @engine=Distribution::T::GSL_
       end
+      it "return correct values for cdf with fractional df" do
+        @engine.cdf(1,2.5).should be_within(1e-7).of(0.7979695)
+      end
     it_should_behave_like "T engine"    
     it_should_behave_like "T engine(with pdf)"    
     end
-  end  
+  end
+=begin
   if Distribution.has_statistics2?
     describe Distribution::T::Statistics2_ do
       before do
@@ -84,7 +102,7 @@ end
     it_should_behave_like "T engine"    
     end  
   end
-  
+=end  
   if Distribution.has_java?
     describe Distribution::T::Java_ do
       before do

metadata

@@ -1,12 +1,8 @@
 --- !ruby/object:Gem::Specification 
 name: distribution
 version: !ruby/object:Gem::Version 
-  prerelease: false
-  segments: 
-  - 0
-  - 4
-  - 0
-  version: 0.4.0
+  prerelease: 
+  version: 0.5.0
 platform: ruby
 authors: 
 - Claudio Bustos
@@ -35,7 +31,7 @@ cert_chain:
   rpP0jjs0
   -----END CERTIFICATE-----
 
-date: 2011-02-01 00:00:00 -03:00
+date: 2011-05-03 00:00:00 -03:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -46,10 +42,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 2
-        - 0
-        - 4
         version: 2.0.4
   type: :development
   version_requirements: *id001
@@ -61,9 +53,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 2
-        - 0
         version: "2.0"
   type: :development
   version_requirements: *id002
@@ -75,8 +64,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 0
         version: "0"
   type: :development
   version_requirements: *id003
@@ -88,10 +75,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 2
-        - 8
-        - 0
         version: 2.8.0
   type: :development
   version_requirements: *id004
@@ -124,6 +107,7 @@ files:
 - benchmark/factorial_hash.rb
 - benchmark/factorial_method.rb
 - benchmark/odd.rb
+- benchmark/power.rb
 - bin/distribution
 - data/template/distribution.erb
 - data/template/distribution/gsl.erb
@@ -156,6 +140,8 @@ files:
 - lib/distribution/hypergeometric/gsl.rb
 - lib/distribution/hypergeometric/java.rb
 - lib/distribution/hypergeometric/ruby.rb
+- lib/distribution/logistic.rb
+- lib/distribution/logistic/ruby.rb
 - lib/distribution/math_extension.rb
 - lib/distribution/normal.rb
 - lib/distribution/normal/gsl.rb
@@ -178,6 +164,7 @@ files:
 - spec/exponential_spec.rb
 - spec/f_spec.rb
 - spec/hypergeometric_spec.rb
+- spec/logistic_spec.rb
 - spec/math_extension_spec.rb
 - spec/normal_spec.rb
 - spec/poisson_spec.rb
@@ -200,21 +187,17 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      segments: 
-      - 0
       version: "0"
 required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      segments: 
-      - 0
       version: "0"
 requirements: []
 
 rubyforge_project: distribution
-rubygems_version: 1.3.7
+rubygems_version: 1.6.0
 signing_key: 
 specification_version: 3
 summary: Statistical Distributions library