summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorAndreas Maunz <andreas@maunz.de>2012-04-13 16:10:23 +0200
committerAndreas Maunz <andreas@maunz.de>2012-04-13 16:10:23 +0200
commit612a1bf1ab41ed366486d69153e8e4c8ea39e576 (patch)
tree8042ea004b68dbdeb8ca33e543250a44cfd9de4e
parentb78bda6de028bcf93304353d1ffa06326d1ee6e6 (diff)
Added min_sim transform to cosine space
-rw-r--r--lazar.rb4
1 files changed, 3 insertions, 1 deletions
diff --git a/lazar.rb b/lazar.rb
index 51a49aa..68878f7 100644
--- a/lazar.rb
+++ b/lazar.rb
@@ -125,7 +125,9 @@ post '/lazar/?' do
else
lazar.similarity_algorithm = "Similarity.cosine"
- min_sim = 0.4 unless params[:min_sim]
+ min_sim = 0.7 unless params[:min_sim]
+ min_sim = (min_sim * 2.0 -1.0 ) # transform to cosine range [-1,1]
+
training_features_tl = training_features.features.collect{|f,info| info[DC.description].gsub(/.*\[/,"").chop.split(", ")}
training_features_pc_types = training_features_tl.collect{|info| info[0]}.flatten.uniq
training_features_lib = training_features_tl.collect{|info| info[1]}.flatten.uniq