1 files changed, 23 insertions, 0 deletions
diff --git a/paper/crossvalidation.rb b/paper/crossvalidation.rb
new file mode 100644
index 0000000..1f4c023
--- /dev/null
+++ b/paper/crossvalidation.rb
@@ -0,0 +1,23 @@
+require_relative 'include.rb'
+
+name = File.basename ARGV[0], ".csv"
+file = File.join DATA,ARGV[0]
+csv_file = File.join(DATA,ARGV[0].sub(/.csv/,'-cv.csv'))
+id_file = File.join(DATA,ARGV[0].sub(/.csv/,'-cv.id'))
+dataset = Dataset.from_csv_file file
+model = Model::LazarRegression.create dataset
+cv = RegressionCrossValidation.create model
+File.open(id_file,"w+"){|f| f.puts cv.id}
+
+data = []
+cv.predictions.each do |p|
+  smi = Compound.find(p[0]).smiles
+  data << [smi,p[1].median,p[2],p[3]]
+end
+
+data.sort!{|a,b| a[1] <=> b[1]}
+
+CSV.open(csv_file,"w+") do |csv|
+  csv << ["SMILES","LOAEL_measured_median","LOAEL_predicted","Confidence"]
+  data.each{|r| csv << r}
+end