def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=1,
stratify=y)
y_train_01_subset = y_train.copy()
y_train_02_subset = y_train.copy()
y_train_01_subset[(y_train == 1) | (y_train == 2)] = -1
y_train_01_subset[(y_train_01_subset == 0)] = 1
y_train_02_subset[(y_train == 1) | (y_train == 0)] = -1
y_train_02_subset[(y_train_02_subset == 2)] = 1
ppn1 = Perceptron()
ppn2 = Perceptron()
ppn1.fit(X_train, y_train_01_subset)
ppn2.fit(X_train, y_train_02_subset)
classifier = Classifier(ppn1, ppn2)
plot_decision_regions(X_train, y_train, classifier=classifier)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=1,
stratify=y)
#w regresji logarytmicznej wyjście przyjmuje wartości 0 lub 1 (prawdopodobieństwa)
X_train_01_subset = X_train[(y_train == 0) | (y_train == 1)]
y_train_01_subset = y_train[(y_train == 0) | (y_train == 1)]
lrgd = LogisticRegressionGD(eta=0.05, n_iter=1000, random_state=1)
lrgd.fit(X_train_01_subset, y_train_01_subset)
plot_decision_regions(X=X_train_01_subset,
y=y_train_01_subset,
classifier=lrgd)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=1,
stratify=y)
X_train_01_subset = X_train[(y_train == 0) | (y_train == 1)]
y_train_01_subset = y_train[(y_train == 0) | (y_train == 1)]
# w perceptronie wyjście jest albo 1 albo -1
y_train_01_subset[(y_train_01_subset == 0)] = -1
ppn = Perceptron(eta=0.1, n_iter=10)
ppn.fit(X_train_01_subset, y_train_01_subset)
plot_decision_regions(X=X_train_01_subset,
y=y_train_01_subset,
classifier=ppn)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=1, stratify=y)
#w regresji logarytmicznej wyjście przyjmuje wartości 0 lub 1 (prawdopodobieństwa)
y_train_01_subset = y_train.copy()
y_train_02_subset = y_train.copy()
y_train_01_subset[(y_train == 0) | (y_train == 1)] = 1
y_train_01_subset[(y_train_01_subset == 2)] = 0
y_train_02_subset[(y_train == 2) | (y_train == 0)] = 0
y_train_02_subset[(y_train_02_subset == 1)] = 1
lrgd1 = LogisticRegressionGD()
lrgd2 = LogisticRegressionGD()
lrgd1.fit(X_train, y_train_01_subset)
lrgd2.fit(X_train, y_train_02_subset)
classifier = Classifier(lrgd1, lrgd2)
#y_1_activation = classifier.activation(y_train_01_subset)
#y_2_activation = classifier.activation(y_train_02_subset)
y_3_activation = classifier.activation(X_train)
#print(y_1_activation)
#print(y_2_activation)
print(y_3_activation)
plot_decision_regions(X_train, y_train, classifier=classifier)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=1,
stratify=y)
X_train_01_subset, y_train_01_subset, y_train_03_subset = prepare_class_subsets(
X_train, y_train)
print('01_subset ', y_train_01_subset)
print('03_subset ', y_train_03_subset)
ppn1 = LogisticRegressionGD(eta=0., n_iter=15000)
ppn1.fit(X_train_01_subset, y_train_01_subset)
ppn2 = LogisticRegressionGD(eta=0.55, n_iter=15000)
ppn2.fit(X_train_01_subset, y_train_03_subset)
calc_accuracy_total(X_train, ppn1, ppn2, y_train, y_train_01_subset,
y_train_03_subset)
# w perceptronie wyjście jest albo 1 albo -1
# y_train_01_subset[(y_train_01_subset == 0)] = -1
clas = Classifier(ppn1, ppn2)
plot_decision_regions(X_train, y_train, classifier=clas)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=1,
stratify=y)
y_train_01_subset = y_train.copy()
y_train_03_subset = y_train.copy()
X_train_01_subset = X_train.copy()
y_train_01_subset[(y_train == 1) | (y_train == 2)] = -1
y_train_01_subset[(y_train_01_subset == 0)] = 1
y_train_03_subset[(y_train == 1) | (y_train == 0)] = -1
y_train_03_subset[(y_train_03_subset == 2)] = 1
print('y_train_01_subset ', y_train_01_subset)
print('y_train_03_subset ', y_train_03_subset)
ppn1 = Perceptron(eta=0.1, n_iter=500)
ppn1.fit(X_train_01_subset, y_train_01_subset)
ppn2 = Perceptron(eta=0.1, n_iter=500)
ppn2.fit(X_train_01_subset, y_train_03_subset)
y1_predict = ppn1.predict(X_train)
y3_predict = ppn2.predict(X_train)
accuracy_1 = accuracy(ppn1.predict(X_train), y_train_01_subset)
accuracy_3 = accuracy(ppn2.predict(X_train), y_train_03_subset)
print("acc1", accuracy_1)
print("acc2", accuracy_3)
if accuracy_1 > accuracy_3:
y_results = np.where(y1_predict == 0, 0,
np.where(y3_predict == 1, 2, 1))
else:
y_results = np.where(y3_predict == 0, 2,
np.where(y1_predict == 1, 0, 1))
print("acc_total", accuracy(y_results, y_train))
# w perceptronie wyjście jest albo 1 albo -1
# y_train_01_subset[(y_train_01_subset == 0)] = -1
clas = Classifier(ppn1, ppn2)
plot_decision_regions(X_train, y_train, classifier=clas)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
# print('y = ', y)
x_train, x_test, y_train, y_test = train_test_split(X,
y,
test_size=0.3,
random_state=1,
stratify=y)
y_train_03_subset = y_train.copy()
y_train_01_subset = y_train.copy()
x_train_01_subset = x_train.copy()
clasyficationFactory = None
clas = 'Perceptron'
#clas = 'regLog'
if clas == "Perceptron":
clasyficationFactory = Perceptron()
y_train_01_subset[(y_train == 1) | (y_train == 2)] = -1
y_train_01_subset[(y_train_01_subset == 0)] = 1
y_train_03_subset[(y_train == 1) | (y_train == 0)] = -1
y_train_03_subset[(y_train_03_subset == 2)] = 1
else:
clasyficationFactory = LogisticRegressionGD()
y_train_01_subset[(y_train == 1) | (y_train == 2)] = 1
y_train_01_subset[(y_train_01_subset == 0)] = 0
y_train_03_subset[(y_train == 1) | (y_train == 0)] = 1
y_train_03_subset[(y_train_03_subset == 2)] = 0
ppn1 = clasyficationFactory.startTrain(x_train_01_subset,
y_train_01_subset)
ppn3 = clasyficationFactory.startTrain(x_train_01_subset,
y_train_03_subset)
if clas == 'regLog':
probabilityofLogicReggression(ppn1, ppn3, x_train_01_subset)
classifier = clasyficationFactory.classifierFactory(ppn1, ppn3)
plot_decision_regions(x_train, y_train, classifier=classifier)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
x = iris.data[:, [2, 3]]
y = iris.target
x_train, x_test, y_train, y_test = train_test_split(x,
y,
test_size=0.3,
random_state=1,
stratify=y)
multi_classifier = MultiClassifier(x_train, y_train)
plot_decision_regions(X=x_test, y=y_test, classifier=multi_classifier)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [1, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=1, stratify=y)
multi = Multiclass()
multi.fit(X_train, y_train)
multi.predict(X_test)
plot_decision_regions(X=X_test, y=y_test, classifier=multi)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
print(iris.feature_names)
X = iris.data[:, [1, 2]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=1,
stratify=y)
y_train_logreg1 = y_train.copy()
y_train_logreg2 = y_train.copy()
X_train_logreg = X_train.copy()
y_train_logreg1[(y_train == 1) | (y_train == 2)] = 0
y_train_logreg1[(y_train == 0)] = 1
y_train_logreg2[(y_train == 1) | (y_train == 0)] = 0
y_train_logreg2[(y_train == 2)] = 1
logreg1 = LogisticRegressionGD(eta=0.01)
logreg2 = LogisticRegressionGD(eta=0.01)
logreg1.fit(X_train_logreg, y_train_logreg1)
logreg2.fit(X_train_logreg, y_train_logreg2)
plot_decision_regions_part(X=X_train_logreg,
y=y_train_logreg1,
classifier=logreg1)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
plot_decision_regions_part(X=X_train_logreg,
y=y_train_logreg2,
classifier=logreg2)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
clsif = Classifier(logreg1, logreg2)
# predicted -1's and 1's
logreg1_pred = logreg1.predict(X_train)
logreg2_pred = logreg2.predict(X_train)
# partial accuracies
ac_logreg1 = accuracy_score(logreg1_pred, y_train_logreg1)
ac_logreg2 = accuracy_score(logreg2_pred, y_train_logreg2)
# overall accuracy
if ac_logreg1 > ac_logreg2:
y_res = np.where(logreg1_pred == 1, 0,
np.where(logreg2_pred == 1, 2, 1))
else:
y_res = np.where(logreg2_pred == 1, 2,
np.where(logreg1_pred == 1, 0, 1))
print(f'Overall accuracy:', round(accuracy_score(y_res, y_train), 3))
plot_decision_regions(X=X_train_logreg, y=y_train, classifier=clsif)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.legend(loc='upper left')
plt.show()
def main():
iris = datasets.load_iris()
X = iris.data[:, [2, 3]]
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=1,
stratify=y)
y_train_perc_1 = y_train.copy()
y_train_perc_2 = y_train.copy()
X_train_perc = X_train.copy()
y_train_perc_1[(y_train == 1) | (y_train == 2)] = -1
y_train_perc_1[(y_train == 0)] = 1
y_train_perc_2[(y_train == 1) | (y_train == 0)] = -1
y_train_perc_2[(y_train == 2)] = 1
perc1 = Perceptron(eta=0.1)
perc2 = Perceptron(eta=0.01, n_iter=173)
perc1.fit(X_train_perc, y_train_perc_1)
perc2.fit(X_train_perc, y_train_perc_2)
plot_decision_regions_part(X=X_train_perc,
y=y_train_perc_1,
classifier=perc1)
plt.xlabel(r'sepal width (cm)')
plt.ylabel(r'petal length (cm)')
plt.legend(loc='upper left')
plt.title("Class division by first perceptron")
plt.show()
plot_decision_regions_part(X=X_train_perc,
y=y_train_perc_2,
classifier=perc2)
plt.xlabel(r'sepal width (cm)')
plt.ylabel(r'petal length (cm)')
plt.legend(loc='upper left')
plt.title("Class division by second perceptron")
plt.show()
clsif = Classifier(perc1, perc2)
# predicted -1's and 1's
perc1_pred = perc1.predict(X_train)
perc2_pred = perc2.predict(X_train)
# partial accuracies
ac_perc1 = accuracy_score(perc1_pred, y_train_perc_1)
ac_perc2 = accuracy_score(perc2_pred, y_train_perc_2)
# overall accuracy
if ac_perc1 > ac_perc2:
y_res = np.where(perc1_pred == 1, 0, np.where(perc2_pred == 1, 2, 1))
else:
y_res = np.where(perc2_pred == 1, 2, np.where(perc1_pred == 1, 0, 1))
print(f'Overall accuracy:', round(accuracy_score(y_res, y_train), 3))
print(iris.feature_names)
plot_decision_regions(X=X_train_perc, y=y_train, classifier=clsif)
plt.xlabel(r'sepal width (cm)')
plt.ylabel(r'petal length (cm)')
plt.legend(loc='upper left')
plt.title("Overall class division by multiperceptron model")
plt.show()
onehotencoder = OneHotEncoder(categorical_features=[0]) y_st = np.reshape(y, (-1, 1)) y_st = onehotencoder.fit_transform(y_st).toarray() # w perceptronie wyjście jest albo 1 albo -1 # y_train_01_subset[(y_train_01_subset == 0)] = -1 y_st[y_st == 0] = -1 X_train, X_test, y_train, y_test = train_test_split(X_st, y_st, test_size=0.3, random_state=1, stratify=y) mclass = MultiClassPredict(eta=0.05, n_iter=10, classes=3) mclass.fit(X_train, y_train) predicted = mclass.predict(X_test) y_test[y_test == -1] = 0 y_test = y_test.dot(onehotencoder.active_features_).astype(int) probability = y_test[y_test == predicted].shape[0] / y_test.shape[0] print(probability) from sklearn.metrics import confusion_matrix cnf = confusion_matrix(y_test, predicted) print(cnf) plot_decision_regions(X=X_test, y=y_test, classifier=mclass) plt.xlabel(r'$x_1$') plt.ylabel(r'$x_2$') plt.legend(loc='upper left') plt.show()
def main():
iris = datasets.load_iris()
irisData = iris.data[:, [2, 3]]
irisClass = iris.target
dataTrainingSet, dataTestSet, classTrainingSet, classTestSet = train_test_split(
irisData, irisClass, test_size=0.3, random_state=1, stratify=irisClass)
# =============== Perceptron ====================
# Perceptron 1
classTrainingSubset1 = np.copy(classTrainingSet)
classTrainingSubset1 = classTrainingSubset1[(classTrainingSubset1 != 2)]
dataTrainingSubset1 = np.copy(dataTrainingSet)
dataTrainingSubset1 = dataTrainingSubset1[(classTrainingSet != 2)]
classTrainingSubset1[(classTrainingSubset1 != 0)] = -1
classTrainingSubset1[(classTrainingSubset1 != -1)] = 1
perceptron1 = Perceptron(learningRate=0.1, iterationsToStop=10)
perceptron1.learn(dataTrainingSubset1, classTrainingSubset1)
# Perceptron 2
classTrainingSubset2 = np.copy(classTrainingSet)
classTrainingSubset2 = classTrainingSubset2[(classTrainingSubset2 != 1)]
dataTrainingSubset2 = np.copy(dataTrainingSet)
dataTrainingSubset2 = dataTrainingSubset2[(classTrainingSet != 1)]
classTrainingSubset2[(classTrainingSubset2 != 2)] = -1
classTrainingSubset2[(classTrainingSubset2 != -1)] = 1
perceptron2 = Perceptron(learningRate=0.1, iterationsToStop=10)
perceptron2.learn(dataTrainingSubset2, classTrainingSubset2)
# Perceptron 3
classTrainingSubset3 = np.copy(classTrainingSet)
classTrainingSubset3 = classTrainingSubset3[(classTrainingSubset3 != 0)]
dataTrainingSubset3 = np.copy(dataTrainingSet)
dataTrainingSubset3 = dataTrainingSubset3[(classTrainingSet != 0)]
classTrainingSubset3[(classTrainingSubset3 != 1)] = -1
perceptron3 = Perceptron(learningRate=0.35, iterationsToStop=850)
perceptron3.learn(dataTrainingSubset3, classTrainingSubset3)
multiPerceptron = MultiPerceptron(perceptron1, perceptron2, perceptron3)
plot_decision_regions(X=dataTestSet,
y=classTestSet,
classifier=multiPerceptron)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.title('Perceptron')
plt.legend(loc='upper left')
plt.show()
# =============== Logistic regression ====================
classTrainingSubset1[(classTrainingSubset1 != 1)] = 0
logisticRegression1 = LogisticRegression(learningRate=0.05,
iterationsToStop=1000,
random_state=1)
logisticRegression1.learn(dataTrainingSubset1, classTrainingSubset1)
logisticRegression1.printProbability(dataTrainingSubset1)
classTrainingSubset2[(classTrainingSubset2 != 1)] = 0
logisticRegression2 = LogisticRegression(learningRate=0.05,
iterationsToStop=1000,
random_state=1)
logisticRegression2.learn(dataTrainingSubset2, classTrainingSubset2)
logisticRegression2.printProbability(dataTrainingSubset2)
classTrainingSubset3[(classTrainingSubset3 != 1)] = 0
logisticRegression3 = LogisticRegression(learningRate=0.15,
iterationsToStop=1500,
random_state=1)
logisticRegression3.learn(dataTrainingSubset3, classTrainingSubset3)
logisticRegression3.printProbability(dataTrainingSubset3)
multiLogisticRegression = MultiLogisticRegression(logisticRegression1,
logisticRegression2,
logisticRegression3)
plot_decision_regions(X=dataTestSet,
y=classTestSet,
classifier=multiLogisticRegression)
plt.xlabel(r'$x_1$')
plt.ylabel(r'$x_2$')
plt.title('Logistic regression')
plt.legend(loc='lower right')
plt.show()
