summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorChristoph Helma <helma@in-silico.ch>2015-10-07 12:45:10 +0200
committerChristoph Helma <helma@in-silico.ch>2015-10-07 12:45:10 +0200
commit72e2a77fb0b72f7d525ce4b51fc03c8ee9920da2 (patch)
tree3410deb787e8c6ac6ab00c2cf526eae167b731c5
parent8d2f1c8a0f6cc9f7a481d1117bf8b3351130b1ea (diff)
confidence = similarity to nearest neighbor
-rw-r--r--lib/classification.rb4
-rw-r--r--lib/regression.rb5
2 files changed, 7 insertions, 2 deletions
diff --git a/lib/classification.rb b/lib/classification.rb
index ab1efd8..0a32126 100644
--- a/lib/classification.rb
+++ b/lib/classification.rb
@@ -8,8 +8,10 @@ module OpenTox
return {:value => nil,:confidence => nil,:warning => "Cound not find similar compounds."} if neighbors.empty?
weighted_sum = {}
sim_sum = 0.0
+ confidence = 0.0
neighbors.each do |row|
n,sim,acts = row
+ confidence = sim if sim > confidence # distance to nearest neighbor
acts.each do |act|
weighted_sum[act] ||= 0
weighted_sum[act] += sim
@@ -22,7 +24,7 @@ module OpenTox
sim_sum = weighted_sum[weighted_sum.keys[0]]
sim_sum -= weighted_sum[weighted_sum.keys[1]]
sim_sum > 0 ? prediction = weighted_sum.keys[0] : prediction = weighted_sum.keys[1]
- confidence = (sim_sum/neighbors.size).abs
+ #confidence = (sim_sum/neighbors.size).abs
return {:value => prediction,:confidence => confidence}
else
bad_request_error "Cannot predict more than 2 classes, multinomial classifications is not yet implemented. Received classes were: '#{weighted.sum.keys}'"
diff --git a/lib/regression.rb b/lib/regression.rb
index 2580a1e..9062a9e 100644
--- a/lib/regression.rb
+++ b/lib/regression.rb
@@ -22,15 +22,18 @@ module OpenTox
def self.weighted_average compound, params
weighted_sum = 0.0
sim_sum = 0.0
+ confidence = 0.0
neighbors = params[:neighbors]
neighbors.each do |row|
n,sim,acts = row
+ confidence = sim if sim > confidence # distance to nearest neighbor
+ # TODO add LOO errors
acts.each do |act|
weighted_sum += sim*Math.log10(act)
sim_sum += sim
end
end
- confidence = sim_sum*neighbors.size.to_f/params[:training_dataset_size]
+ #confidence = sim_sum*neighbors.size.to_f/params[:training_dataset_size]
sim_sum == 0 ? prediction = nil : prediction = 10**(weighted_sum/sim_sum)
{:value => prediction,:confidence => confidence}
end