def testTennis(trainFile, testFile, attrFile):
tennis_train = pd.read_csv(trainFile, sep=" ", header=None)
tennis_test = pd.read_csv(testFile, sep=" ", header=None)
tennis_attr = pd.read_csv(attrFile, sep=" ", header=None)
input, output, input_test, output_test = digitalize(tennis_train, tennis_test, tennis_attr,"tennis")
input = np.concatenate((input, output), axis=1)
test_input = np.concatenate((input_test, output_test), axis=1)
# p, numOfRulesPerHypo, numofattr, numofoutput, r, m, fit_threshold, stopGeneration, strategy, dataType):
ga = GA(100, 4, 10, 2, 0.3, 0.1, 1, 1, 30, 'fitness-proportional', 'tennis')
trainAcc, bestHypo = ga.fit(input)
print ('Training data accuracy is: ' + str(trainAcc[0]))
testAcc = ga.predict(bestHypo, test_input)
print ('Test data accuracy is: ' + str(testAcc[0]))
ga.printRules(bestHypo)
def testIris(trainfile, testfile, attrfile):
iris_train = pd.read_csv(trainfile, header=None)
iris_test = pd.read_csv(testfile, header=None)
iris_attr = pd.read_csv(attrfile, header=None)
input, output, input_test, output_test = digitalize(iris_train, iris_test, iris_attr,"iris")
input = input.astype(np.float)
input = np.concatenate((input, output), axis=1)
test_input = np.concatenate((input_test, output_test), axis=1)
# p, numOfRulesPerHypo, numofattr, numofoutput, r, m, numOfChangeBitsPerRule, fit_threshold, stopGeneration, strategy, dataType): tournament fitness-proportional rank
# ga = GA(100, 6, 28, 3, 0.4, 0.2, 1, 0.8, 100, 'rank', 'iris')
# (6*2*4+3)*4 *4
ga = GA(400, 6, 48, 3, 0.3, 0.1, 1, 1, 50, 'tournament', 'iris')
correct, bestHypo = ga.fit(input)
print ('correct rate on traning data' + str(correct))
ga.printRules(bestHypo)
cor = ga.predict(bestHypo, test_input)
