A binary labeling perceptron written in Haskell

Percept.hs

module Percept where

-- The example training/testing data has single digit numbers set to True, and double digit numbers set to False
trainingData :: [([Double], Bool)]
testingData :: [([Double], Bool)]
trainingData = [
	([1], True),
	([5], True),
	([6], True),
	([9], True),
	([15], False),
	([12], False),
	([25], False),
	([19], False)]
testingData = [
	([4], True),
	([2], True),
	([3], True),
	([8], True),
	([13], False),
	([11], False),
	([29], False),
	([18], False)]

learningIterations = 100
startingWeights = take (length $ fst $ testingData !! 0 ) $ repeat 0.0
startingTheta = 0.0

dotProduct x y = sum $ zipWith (*) x y
decision x w theta = dotProduct x w > theta

-- Updates the weights and theta from a single training record.
-- (Training Data, Binary Value of Training Data) -> (Previous Weights, Old Theta) -> (New Weights, New Theta)
perceptron :: ([Double], Double) -> ([Double], Bool) -> ([Double], Double)
perceptron (weights, theta) (trainingData, val)
	| d == False && val == True = (newWeights (+), theta - 1)
	| d == True && val == False = (newWeights (-), theta + 1)
	| otherwise = (weights, theta)
	where
		d = decision trainingData weights theta 
		newWeights combiningFunction = zipWith combiningFunction weights trainingData

-- Loops through a list of training records "learningIterations" times, and returns adjusted weights and theta. 
learn :: ([Double], Double) -> [([Double], Bool)] -> ([Double], Double)
learn (weights, theta) trainingData = last $ take learningIterations $ iterate (\ currentWeightTheta -> foldl perceptron currentWeightTheta trainingData) (startingWeights, startingTheta) 

-- Loops through a list of training records and prints "Correct" if the given weights/theta correctly classified the value, and "Incorrect" otherwise. 
act :: ([Double], Double) -> [([Double], Bool)] -> IO ()
act (weights, theta) testingData = mapM_ (\ (dataSet, val) -> putStrLn $ boolToEnglish $ val == decision dataSet weights theta) testingData
	where
		boolToEnglish True = "Correct"
		boolToEnglish False = "Incorrect"

main :: IO ()
main = (\ x -> act x testingData) $ learn (startingWeights, startingTheta) trainingData