initial refactored version for mutagenicity paper

1 files changed, 71 insertions, 0 deletions

diff --git a/lib/statistics.rb b/lib/statistics.rb
new file mode 100644
index 0000000..15ea416
--- /dev/null
+++ b/lib/statistics.rb
@@ -0,0 +1,71 @@
+class ClassificationStatistics
+
+  def initialize dir
+    @dir = dir
+    @folds = Dir[File.join(@dir,"[0-9]*")]
+    @confusion_matrix_dir = File.join(@dir,"confusion_matrices")
+    @summaries_dir = File.join(@dir,"summaries")
+  end
+
+  def confusion_matrix
+
+    confusion_matrices = {
+      :all => {:tp => 0, :fp => 0, :tn => 0, :fn => 0},
+      :high_confidence  => {:tp => 0, :fp => 0, :tn => 0, :fn => 0},
+      :low_confidence  => {:tp => 0, :fp => 0, :tn => 0, :fn => 0},
+    }
+
+    @folds.each do |dir|
+      test_dir = File.join(dir,"test")
+      classifications = File.readlines(File.join(test_dir,"classification")).collect{|row| row.chomp.split(",")}
+      measurements = File.readlines(File.join(test_dir,"dependent_variables")).collect{|v| v.to_i}
+      similarity_thresholds = File.readlines(File.join(dir,"train","similarity_thresholds")).collect{|v| v.chomp.to_f}
+      classifications.each_with_index do |c,i|
+        prediction = c[1]
+        max_sim = c[4].to_f
+        unless prediction.empty?
+          prediction = prediction.to_i
+          if prediction == 1 and measurements[i] == 1
+            confusion_matrices[:all][:tp] +=1
+            max_sim > similarity_thresholds[1] ?  confusion_matrices[:high_confidence][:tp] +=1 : confusion_matrices[:low_confidence][:tp] +=1
+          elsif prediction == 0 and measurements[i] == 0
+            confusion_matrices[:all][:tn] +=1
+            max_sim > similarity_thresholds[1] ?  confusion_matrices[:high_confidence][:tn] +=1 : confusion_matrices[:low_confidence][:tn] +=1
+          elsif prediction == 1 and measurements[i] == 0
+            confusion_matrices[:all][:fp] +=1
+            max_sim > similarity_thresholds[1] ?  confusion_matrices[:high_confidence][:fp] +=1 : confusion_matrices[:low_confidence][:fp] +=1
+          elsif prediction == 0 and measurements[i] == 1
+            confusion_matrices[:all][:fn] +=1
+            max_sim > similarity_thresholds[1] ?  confusion_matrices[:high_confidence][:fn] +=1 : confusion_matrices[:low_confidence][:fn] +=1
+          end
+        end
+      end
+      FileUtils.mkdir_p @confusion_matrix_dir
+      confusion_matrices.each do |t,m|
+        File.open(File.join(@confusion_matrix_dir,t.to_s),"w+"){ |f| f.puts "#{m[:tp]},#{m[:fp]}\n#{m[:fn]},#{m[:tn]}" }
+      end
+    end
+
+  end
+
+  def summary
+    [:all,:high_confidence,:low_confidence].each do |cat|
+      confusion_matrix_file = File.join(@confusion_matrix_dir,cat.to_s)
+      confusion_matrix unless File.exists? confusion_matrix_file
+      matrix = File.readlines(confusion_matrix_file).collect{|row| row.chomp.split(",").collect{|v| v.to_f}}
+      tp = matrix[0][0]
+      fp = matrix[0][1]
+      fn = matrix[1][0]
+      tn = matrix[1][1]
+      FileUtils.mkdir_p @summaries_dir
+      File.open(File.join(@summaries_dir,cat.to_s),"w+") do |f|
+        f.puts "accuracy,#{(tp+tn)/(tp+fp+tn+fn)}"
+        f.puts "true_positive_rate,#{tp/(tp+fn)}"
+        f.puts "true_negative_rate,#{tn/(tn+fp)}"
+        f.puts "positive_predictive_value,#{tp/(tp+fp)}"
+        f.puts "negative_predictive_value,#{tn/(tn+fn)}"
+      end
+    end
+  end
+end
+