Helper functions for confidence bounds

src/Data/Conduit/Bootstrap.hs

@@ -3,7 +3,7 @@
 -- Bootstrap re-sampling
 
 module Data.Conduit.Bootstrap
-       (bootstrapC)
+       (bootstrapC, getConfidenceBounds, defaultConfidenceLevel)
        where
 
 import Data.Conduit
@@ -12,8 +12,7 @@ import qualified Data.Conduit.Combinators as CC
 import Control.Monad.Trans.Resource
 import Data.Vector
 import qualified Data.Vector.Generic as VG
-
+import Data.List (sort)
-import Debug.Trace
 
 import System.Random (mkStdGen, randomRs)
 
@@ -26,3 +25,15 @@ bootstrapC numberOfSamples final = do
 
 resampleVector gen v = Prelude.map (\ix -> v VG.! ix) $ Prelude.take n $ randomRs (0, n-1) gen
   where n = VG.length v
+
+defaultConfidenceLevel = 0.95
+
+getConfidenceBounds :: Ord a => Double -> [a] -> (a, a)
+getConfidenceBounds confidenceLevel samples = ((samplesSorted !! toBeCut), (samplesSorted !! (n - 1 - toBeCut)))
+  where n = Prelude.length samples
+        toBeCut' = floor (((1 - confidenceLevel + epsilon) * (fromIntegral n)) / 2)
+        toBeCut = if 2 * toBeCut' >= n
+                  then toBeCut' - 1
+                  else toBeCut'
+        samplesSorted = sort samples
+        epsilon = 0.0001

test/Spec.hs

@@ -553,6 +553,11 @@ main = hspec $ do
       results <- runResourceT $ runConduit (CL.sourceList listChecked .| bootstrapC nbOfSamples CC.product)
       Prelude.length results `shouldBe` nbOfSamples
       (Prelude.length (Prelude.filter (> 0) results)) `shouldNotBe` 0
+    it "test gettings bounds" $ do
+      let sample = [3.0, 11.0, 2.0, 4.0, 15.0,  12.0, 2013.5, 19.0, 17.0, -10000.0,
+                    16.0, 13.0, 6.0, 7.0, 8.0,  5.0, 9.0, 10.0, 14.0, 18]
+      getConfidenceBounds defaultConfidenceLevel sample `shouldBe` (-10000.0, 2013.5)
+      getConfidenceBounds 0.9 sample `shouldBe` (2.0, 19.0)
   describe "tokenizer" $ do
     it "simple utterance with '13a' tokenizer" $ do
       tokenize (Just V13a) "To be or not to be, that's the question." `shouldBe`