diff --git a/data/T10I4D100K.dat b/data/T10I4D100K.dat new file mode 100644 # file too large to diff: data/T10I4D100K.dat diff --git a/hinduce-examples.cabal b/hinduce-examples.cabal --- a/hinduce-examples.cabal +++ b/hinduce-examples.cabal @@ -1,5 +1,5 @@ Name: hinduce-examples -Version: 0.0.0.1 +Version: 0.0.0.2 License: BSD3 Author: Robert Hensing Synopsis: Example data for hInduce @@ -13,6 +13,7 @@ data/iris/iris.data data/iris/bezdekIris.data data/iris/iris.names + data/T10I4D100K.dat Library Build-Depends: base >= 4 && < 5 @@ -21,7 +22,14 @@ , csv >= 0.1.2 , hinduce-classifier >= 0.0.0.0 , hinduce-classifier-decisiontree >= 0.0.0.0 + , hinduce-associations-apriori >= 0.0.0.0 , convertible + , containers >= 0.4.2.0 + , vector >= 0.9.1 Exposed-Modules: Data.HInduce.Examples + -- These need not be exposed, but we do to make + -- the haddock documentation more readable. + , Data.HInduce.Examples.DecisionTree + , Data.HInduce.Examples.Associations Other-Modules: Paths_hinduce_examples Hs-Source-Dirs: src diff --git a/src/Data/HInduce/Examples.hs b/src/Data/HInduce/Examples.hs --- a/src/Data/HInduce/Examples.hs +++ b/src/Data/HInduce/Examples.hs @@ -1,89 +1,27 @@ -{-# LANGUAGE MultiParamTypeClasses #-} -module Data.HInduce.Examples ( - -- * Re-exports - module Data.HInduce.Classifier +-- | This package / module provides example data and example code to help you +-- get started with HInduce. You are advised to import this module (ghci: +-- @:m + Data.HInduce.Examples@), not the individual ones below, because +-- @Data.HInduce.Examples@ re-exports modules that are required to run the +-- examples yourself. +-- +-- Click on a module below to view the examples you're interested in. Also note +-- the grey "Source" links at the right hand site of the webpage. +module Data.HInduce.Examples ( -- * The Examples + module Data.HInduce.Examples.DecisionTree + , module Data.HInduce.Examples.Associations + -- * Re-exports for convenience + , module Data.HInduce.Classifier , module Data.HInduce.Classifier.DecisionTree + , module Data.HInduce.Associations.Apriori , module Data.List.HIUtils , module Text.Layout - , module Data.Convertible - -- * Helpers (TODO move to module) - , readCSV - -- * Iris data set - -- | Taken from the UCI Machine Learning Repository: - -- - -- Let's build a decision tree and try it: - -- - -- >>> let model = buildDTree (genMany autoDeciders) irisAttrs irisClass iris - -- - -- >>> classify model [5,4,2,1] - -- Setosa - -- >>> iris !! 10 - -- Iris {sepalLength = 5.4, sepalWidth = 3.7, petalLength = 1.5, petalWidth = 0.2, irisClass = Setosa} - -- - -- Seems good! But can we really know that? - -- Let's train and test on separate data - -- - -- >>> let model' = buildDTree (genMany autoDeciders) irisAttrs irisClass (oddIx iris) - -- - -- >>> dt $ confusion' model' (map (irisAttrs &&& irisClass) $ evenIx iris) - -- Table: Confusion Matrix - -- ||-->Actual - -- Predicted\/ Setosa Versicolor Virginica - -- Setosa 0.3333333333333333 - -- Versicolor 0.30666666666666664 4.0e-2 - -- Virginica 2.666666666666667e-2 0.29333333333333333 - -- - -- Now we see that even though not the whole data set was available - -- when the model was induced, only few misclassifications occur. - - , Iris(..), IrisClass(..), irisAttrs, irisAttrs', readIris, iris - ) where -import Paths_hinduce_examples + ) where +import Data.HInduce.Examples.DecisionTree +import Data.HInduce.Examples.Associations import Data.HInduce.Classifier import Data.HInduce.Classifier.DecisionTree -import Text.Layout -import Data.Convertible +import Data.HInduce.Associations.Apriori import Data.List.HIUtils -import System.IO -import Text.CSV -import System.IO.Unsafe - -test :: FilePath -> IO FilePath -test = getDataFileName - -readCSV x = do - f <- getDataFileName $ "data/" ++ x - parseCSVFromFile f - -openDataR x = do - f <- getDataFileName $ "data/" ++ x - openFile f ReadMode - -data IrisClass = Setosa | Versicolor | Virginica - deriving (Eq, Ord, Show, Read) -instance Layout IrisClass DisplayText where format = fromShow - -data Iris = Iris { sepalLength :: Double - , sepalWidth :: Double - , petalLength :: Double - , petalWidth :: Double - , irisClass :: IrisClass - } - deriving (Eq, Ord, Show, Read) - -irisAttrs (Iris p q r s _) = [p, q, r, s] -irisAttrs' (Iris p q r s _) = ((p, q), (r, s)) - -readIris = do - (Right csv) <- readCSV "iris/iris.data" - return $ map readIrisEntry $ filter (/=[""]) csv - -readIrisEntry [p,q,r,s,"Iris-setosa"] = - Iris (read p) (read q) (read r) (read s) Setosa -readIrisEntry [p,q,r,s,"Iris-versicolor"] = - Iris (read p) (read q) (read r) (read s) Versicolor -readIrisEntry [p,q,r,s,"Iris-virginica"] = - Iris (read p) (read q) (read r) (read s) Virginica +import Text.Layout -iris = unsafePerformIO readIris diff --git a/src/Data/HInduce/Examples/Associations.hs b/src/Data/HInduce/Examples/Associations.hs new file mode 100644 --- /dev/null +++ b/src/Data/HInduce/Examples/Associations.hs @@ -0,0 +1,27 @@ +-- | An example of association rule mining: +-- +-- >>> rules transactions items (top ((take 40) . (filter (\(_,a)->a>= 60)))) +-- > fromList [((fromList [32],fromList [947]),0.1694915254237288),((fromList [39],fromList [145]),0.17238139971817754),((fromList [39],fromList [145,419]),8.266791921089714e-2),((fromList [39],fromList [368]),0.1326914044152184),((fromList [39],fromList [419]),0.1200093940817285), +module Data.HInduce.Examples.Associations where +import Data.HInduce.Associations.Apriori +import Text.Layout +import System.IO.Unsafe +import qualified Data.Set as S +import qualified Data.Vector as V +import Control.Arrow +import Data.Set +import Data.Vector + +import Paths_hinduce_examples + +-- | The transactions in the T10I4D100K.dat data set. +transactions :: Data.Vector.Vector (Data.Set.Set Int) + +-- | The items in the T10I4D100K.dat data set. +items :: Data.Set.Set Int + +(transactions, items) = unsafePerformIO $ do + name <- getDataFileName "data/T10I4D100K.dat" + ds <- loadDataSet name + return $ V.fromList &&& S.unions $ ds + diff --git a/src/Data/HInduce/Examples/DecisionTree.hs b/src/Data/HInduce/Examples/DecisionTree.hs new file mode 100644 --- /dev/null +++ b/src/Data/HInduce/Examples/DecisionTree.hs @@ -0,0 +1,80 @@ +{-# LANGUAGE MultiParamTypeClasses #-} +module Data.HInduce.Examples.DecisionTree ( + -- * Helpers (TODO move to module) + readCSV + -- * Iris data set + -- | Taken from the UCI Machine Learning Repository: + -- + -- Let's build a decision tree and try it: + -- + -- >>> let model = buildDTree (genMany autoDeciders) irisAttrs irisClass iris + -- + -- >>> classify model [5,4,2,1] + -- Setosa + -- >>> iris !! 10 + -- Iris {sepalLength = 5.4, sepalWidth = 3.7, petalLength = 1.5, petalWidth = 0.2, irisClass = Setosa} + -- + -- Seems good! But can we really know that? + -- Let's train and test on separate data + -- + -- >>> let model' = buildDTree (genMany autoDeciders) irisAttrs irisClass (oddIx iris) + -- + -- >>> dt $ confusion' model' (map (irisAttrs &&& irisClass) $ evenIx iris) + -- Table: Confusion Matrix + -- ||-->Actual + -- Predicted\/ Setosa Versicolor Virginica + -- Setosa 0.3333333333333333 + -- Versicolor 0.30666666666666664 4.0e-2 + -- Virginica 2.666666666666667e-2 0.29333333333333333 + -- + -- Now we see that even though not the whole data set was available + -- when the model was induced, only few misclassifications occur. + + , Iris(..), IrisClass(..), irisAttrs, irisAttrs', readIris, iris + ) where +import Paths_hinduce_examples + +import Data.HInduce.Classifier +import Data.HInduce.Classifier.DecisionTree +import Text.Layout +import Data.Convertible +import Data.List.HIUtils +import System.IO +import Text.CSV +import System.IO.Unsafe + +readCSV x = do + f <- getDataFileName $ "data/" ++ x + parseCSVFromFile f + +openDataR x = do + f <- getDataFileName $ "data/" ++ x + openFile f ReadMode + +data IrisClass = Setosa | Versicolor | Virginica + deriving (Eq, Ord, Show, Read) +instance Layout IrisClass DisplayText where format = fromShow + +data Iris = Iris { sepalLength :: Double + , sepalWidth :: Double + , petalLength :: Double + , petalWidth :: Double + , irisClass :: IrisClass + } + deriving (Eq, Ord, Show, Read) + +irisAttrs (Iris p q r s _) = [p, q, r, s] +irisAttrs' (Iris p q r s _) = ((p, q), (r, s)) + +readIris = do + (Right csv) <- readCSV "iris/iris.data" + return $ map readIrisEntry $ filter (/=[""]) csv + +readIrisEntry [p,q,r,s,"Iris-setosa"] = + Iris (read p) (read q) (read r) (read s) Setosa +readIrisEntry [p,q,r,s,"Iris-versicolor"] = + Iris (read p) (read q) (read r) (read s) Versicolor +readIrisEntry [p,q,r,s,"Iris-virginica"] = + Iris (read p) (read q) (read r) (read s) Virginica + +iris = unsafePerformIO readIris