def statistics(trainingData, testData, learner, protectedIndex, protectedValue): h = learner(trainingData) trainingError = labelError(trainingData, h) testError = labelError(testData, h) sp = statisticalParity(testData, protectedIndex, protectedValue, h) UBIF = individualFairness(trainingData, learner, FLIP_PROPORTION) return trainingError, testError, sp, UBIF
def runBaseline(trainingData, testData, learner, protectedIndex, protectedValue):
h = learner(trainingData)
print("Training error: %f" % labelError(trainingData, h))
print("Test error: %f" % labelError(testData, h))
print("SP of the hypothesis: %f" % statisticalParity(testData, protectedIndex, protectedValue, h))
UBIF = individualFairness(trainingData, learner, FLIP_PROPORTION)
print("UBIF of the hypothesis on training: %f" % UBIF)
def statistics(trainingData, testData, learner, protectedIndex,
protectedValue):
h = learner(trainingData)
trainingError = labelError(trainingData, h)
testError = labelError(testData, h)
sp = statisticalParity(testData, protectedIndex, protectedValue, h)
UBIF = individualFairness(trainingData, learner, FLIP_PROPORTION)
return trainingError, testError, sp, UBIF
def runBaseline(trainingData, testData, learner, protectedIndex,
protectedValue):
h = learner(trainingData)
print("Training error: %f" % labelError(trainingData, h))
print("Test error: %f" % labelError(testData, h))
print("SP of the hypothesis: %f" %
statisticalParity(testData, protectedIndex, protectedValue, h))
UBIF = individualFairness(trainingData, learner, FLIP_PROPORTION)
print("UBIF of the hypothesis on training: %f" % UBIF)
def statistics(train, test, protectedIndex, protectedValue, learner):
h = learner(train, protectedIndex, protectedValue)
print("Computing error")
error = labelError(test, h)
print("Computing bias")
bias = signedStatisticalParity(test, protectedIndex, protectedValue, h)
print("Computing UBIF")
ubif = individualFairness(train, learner, 0.2, passProtected=True)
return error, bias, ubif
def statistics(train, test, protectedIndex, protectedValue, learner):
h = learner(train, protectedIndex, protectedValue)
print("Computing error")
error = labelError(test, h)
print("Computing bias")
bias = signedStatisticalParity(test, protectedIndex, protectedValue, h)
print("Computing UBIF")
ubif = individualFairness(train, learner, flipProportion=0.2, passProtected=True)
return error, bias, ubif
def statistics(massager, trainingData, testData, protectedIndex, protectedValue,
learner, flipProportion=0.2):
massagedData = massager(trainingData, protectedIndex, protectedValue)
h = learner(massagedData)
error = labelError(testData, h)
bias = signedStatisticalParity(testData, protectedIndex, protectedValue, h)
ubif = individualFairness(trainingData, learner, flipProportion)
return error, bias, ubif
def statistics(train, test, protectedIndex, protectedValue, numRounds=20): weight = 0.5 flipProportion = 0.2 error = makeErrorFunction(protectedIndex, protectedValue, weight) weakLearner = lambda draw: buildDecisionStump(draw, errorFunction=error) h = boosting.boost(train, weakLearner=weakLearner) bias = ef.signedStatisticalParity(test, protectedIndex, protectedValue, h) error = ef.labelError(test, h) ubif = ef.individualFairness(train, boosting.boost, flipProportion) return error, bias, ubif
def statistics(train, test, protectedIndex, protectedValue, numRounds=20): weight = 0.5 flipProportion = 0.2 error = makeErrorFunction(protectedIndex, protectedValue, weight) weakLearner = lambda draw: buildDecisionStump(draw, errorFunction=error) h = boosting.boost(train, weakLearner = weakLearner) bias = ef.signedStatisticalParity(test, protectedIndex, protectedValue, h) error = ef.labelError(test, h) ubif = ef.individualFairness(train, boosting.boost, flipProportion) return error, bias, ubif