naive_bayes 0.0.0 → 0.0.1

data/.gitignore

@@ -18,4 +18,4 @@ coverage
 rdoc
 pkg
 
-## PROJECT::SPECIFIC
+## PROJECT::SPECIFIC

data/README.rdoc

@@ -21,6 +21,31 @@ This is an extremely simple, straight forward Naive Bayes implementation.
   
   a.classify(*b)
     #=> [:spam, 0.03125]
+    
+You can also tell your classifier to save itself, so its easy for you to pick up where you left off:
+
+  require 'rubygems'
+  require 'naive_bayes'
+
+  a = NaiveBayes.new(:spam, :ham)
+  a.db_filepath = 'path/to/anywhere.nb'
+  
+  # Save is called after every train function
+  a.train(:spam, 'bad', 'word')
+  a.train(:ham, 'good', 'word')
+
+Some time goes past and we want to classify a new document we just received...
+  
+  require 'rubygems'
+  require 'naive_bayes'
+  
+  a = NaiveBayes.load('path/to/file') 
+  
+  b = "this is a bad sentence".split(' ')
+  
+  # It's as if we were never apart
+  a.classify(*b)
+    #=> [:spam, 0.03125]
   
 == Copyright
 

data/VERSION

@@ -1 +1 @@
-0.0.0
+0.0.1

data/lib/naive_bayes.rb

@@ -13,7 +13,19 @@
 
 class NaiveBayes
   
-  attr_reader :features_count, :klass_count
+  class << self
+    def load(db_path)
+      data = ""
+      File.open(db_path) do |f|
+        while line = f.gets
+          data << line
+        end
+      end
+      Marshal.load(data)
+    end
+  end
+    
+  attr_accessor :db_filepath  
   
   def initialize(*klasses)
     @features_count = {}
@@ -31,6 +43,7 @@ class NaiveBayes
       @features_count[klass][feature] += 1
     end
     @klass_count[klass] += 1
+    save
   end
   
   #P(Class | Item) = P(Item | Class) * P(Class)
@@ -44,6 +57,14 @@ class NaiveBayes
   
   private
   
+  def save
+    if @db_filepath
+      File.open(@db_filepath, "w+") do |f|
+        f.write(Marshal.dump(self))
+      end
+    end
+  end
+  
   # P(Item | Class)
   def prob_of_item_given_a_class(features, klass)
     a = features.inject(1.0) do |sum, feature|

data/naive_bayes.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{naive_bayes}
-  s.version = "0.0.0"
+  s.version = "0.0.1"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["reddavis"]
-  s.date = %q{2009-11-14}
+  s.date = %q{2010-01-05}
   s.description = %q{Simple straight forward Naive Bayes classifier implementation}
   s.email = %q{reddavis@gmail.com}
   s.extra_rdoc_files = [
@@ -25,6 +25,7 @@ Gem::Specification.new do |s|
      "VERSION",
      "lib/naive_bayes.rb",
      "naive_bayes.gemspec",
+     "spec/db/naive.nb",
      "spec/naive_bayes_spec.rb",
      "spec/spec.opts",
      "spec/spec_helper.rb"

data/spec/db/naive.nb

@@ -0,0 +1 @@
+o:NaiveBayes	:@klass_count{:	spamf1:hamf1:@db_filepath"D/Users/reddavis/Documents/projects/naive_bayes/spec/db/naive.nb:@features_count{;}"	wordf1"badf1f0;}"wef1"badf1@:

data/spec/naive_bayes_spec.rb

@@ -1,44 +1,57 @@
 require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
 
 describe "NaiveBayes" do
-  
-  describe "Initialization" do
-    before(:all) do
-      @classifier = NaiveBayes.new(:spam, :ham)
+    
+  describe "Classification" do
+    before do
+      @classifier = create_and_train_classifier
     end
     
-    it "should create a features count for each class" do
-      @classifier.features_count.size.should == 2
+    it "should classify as spam with a score of 0.5" do
+      a = @classifier.classify('bad', 'word')
+      a[0].should == :spam
+      a[1].should == 0.5
     end
   end
   
-  describe "Training" do
-    before(:all) do
+  describe "Save" do
+    before do
       @classifier = NaiveBayes.new(:spam, :ham)
-      @classifier.train(:spam, 'bad', 'word')
-    end
-        
-    it "should train" do
-      @classifier.features_count[:spam].size.should == 2
+      @classifier.db_filepath = db_filepath
     end
     
-    it "should bump klass_count for spam up to 1" do
-      @classifier.klass_count[:spam].should == 1
+    it "should save to the filepath provided" do
+      FileUtils.rm(db_filepath, :force => true)
+      @classifier.train(:spam, 'bad')
+      File.exists?(db_filepath).should be_true
     end
   end
   
-  describe "Classification" do
+  describe "Load" do
     before do
-      @classifier = NaiveBayes.new(:spam, :ham)
-      @classifier.train(:spam, 'bad', 'word')
-      @classifier.train(:ham, 'we', 'bad')
+      classifier = NaiveBayes.new(:spam, :ham)
+      classifier.db_filepath = db_filepath
+      classifier.train(:spam, 'bad', 'word')
+      classifier.train(:ham, 'we', 'bad')
     end
     
-    it "should" do
-      a = @classifier.classify('bad', 'word')
-      a[0].should == :spam
-      a[1].should == 0.5
+    it "should return 0.5" do
+      classifier = NaiveBayes.load(db_filepath)
+      classifier.classify('bad', 'word')[1].should == 0.5
     end
   end
   
+  private
+  
+  def create_and_train_classifier
+    a = NaiveBayes.new(:spam, :ham)
+    a.train(:spam, 'bad', 'word')
+    a.train(:ham, 'we', 'bad')
+    a
+  end
+  
+  def db_filepath
+    File.expand_path(File.dirname(__FILE__) + '/db/naive.nb')
+  end
+  
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: naive_bayes
 version: !ruby/object:Gem::Version 
-  version: 0.0.0
+  version: 0.0.1
 platform: ruby
 authors: 
 - reddavis
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-11-14 00:00:00 +00:00
+date: 2010-01-05 00:00:00 +00:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -40,6 +40,7 @@ files:
 - VERSION
 - lib/naive_bayes.rb
 - naive_bayes.gemspec
+- spec/db/naive.nb
 - spec/naive_bayes_spec.rb
 - spec/spec.opts
 - spec/spec_helper.rb