def bagging(x_in, y_in, x_test_in, y_test_in):
"""改变权重,训练多个分类器,投票"""
# 先预测
clf1 = LogisticRegression().fit(x_in, y_in)
predict1 = clf1.predict(x_test_in, y_test_in)
clf2 = LinearSVM().fit(x_in, y_in)
predict2 = clf2.predict(x_test_in, y_test_in)
clf3 = CartDecisionTree().fit(x_in, y_in)
predict3 = clf3.predict(x_test_in, y_test_in)
# 收集投票
predict = np.zeros_like(predict1)
count = 0
for i in range(np.size(y_test_in, axis=0)):
if predict1[i] == predict2[i]:
predict[i] = predict2[i]
elif predict1[i] == predict3[i]:
predict[i] = predict1[i]
else:
predict[i] = predict2[i]
if predict[i] == y_test_in[i]:
count += 1
acc = count / np.size(y_test_in, axis=0) * 100
print("Bagging ACC: %.2f%%" % acc)
return 0
y = pd.Series(y)
result=[]
x = []
for j in range(30):
ans = []
for i in range(5):
X1 = X[0:i*20]
X2 = X[(i+1)*20:]
X_train = X1.append(X2)
X_test = X[i*20:(i+1)*20]
y1 = y[0:i*20]
y2 = y[(i+1)*20:]
y_train = y1.append(y2)
y_test = y[i*20:(i+1)*20]
clf = LogisticRegression(l1_coef= j*2).l1_fit(X_train,y_train)
y_hat = clf.predict(X_test)
y_t = list(y_test)
answer = 0
for i in range(len(y_t)):
if int(y_hat[i]) == y_t[i]:
answer+=1
ans.append((answer)/len(y_t))
result.append((sum(ans)/len(ans)))
x.append(j*5)
print(ans)
print(result)
plt.plot(x,result)
plt.xlabel("Panelty Coefficient")
plt.ylabel("Accuracy")
plt.show()
from Logistic_Regression import LogisticRegression
from sklearn.datasets import load_breast_cancer
from sklearn.datasets import make_classification
import pandas as pd
from sklearn.model_selection import train_test_split
import numpy as np
X, y = make_classification(n_samples=1000)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=42)
clf = LogisticRegression().l1_fit(X_train, y_train)
y_hat = (clf.predict(X_test))
ans = clf.score1(y_hat,y_test)
print("For L1 regularised Logistic Regression ")
print(ans)
clf = LogisticRegression().l2_fit(X_train, y_train)
y_hat = (clf.predict(X_test))
ans = clf.score2(y_hat,y_test)
print("For L2 regularised Logistic Regression ")
print(ans)
from Logistic_Regression import LogisticRegression
from sklearn.datasets import load_breast_cancer
import pandas as pd
import numpy as np
N = 50
P = 8
X = pd.DataFrame(np.random.randn(N, P))
y = pd.Series(np.random.randint(0,2,N))
clf = LogisticRegression().fit(X, y)
y_hat = (clf.predict(X))
ans = clf.score1(y_hat,y)
print("Accuracy with Gradient_Descent Normally")
print(ans)
clf = LogisticRegression().fit_autograd(X, y)
y_hat = (clf.predict(X))
ans = clf.score2(y_hat,y)
print("Accuracy with Autograd Implementation")
print(ans)
from Logistic_Regression import LogisticRegression
from sklearn.datasets import load_breast_cancer
import pandas as pd
from sklearn.model_selection import KFold
import numpy as np
data = np.array(load_breast_cancer().data)
y = np.array(load_breast_cancer().target)
kf = KFold(n_splits=3)
for train_index,test_index in kf.split(data):
X_train,X_test = data[train_index], data[test_index]
y_train,y_test = y[train_index], y[test_index]
X_train = pd.DataFrame(X_train)
X_test = pd.DataFrame(X_test)
y_train = pd.Series(y_train)
y_test = pd.Series(y_test)
clf = LogisticRegression().fit(X_train, y_train)
y_hat = list(clf.predict(X_test))
y_t = list(y_test)
print("Overall Accuracy with K = 3 Folds")
print(clf.score1(y_hat,y_t))
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.25, random_state = 0)
# Feature Scaling
from sklearn.preprocessing import StandardScaler
sc = StandardScaler()
X_train = sc.fit_transform(X_train)
X_test = sc.transform(X_test)
# Fitting Logistic Regression to the Training set
classifier = LogisticRegression(lr=0.001)
classifier.fit(X_train, y_train)
# Predicting the Test set results
y_pred = np.array(classifier.predict(X_test))
# Making the Confusion Matrix
from sklearn.metrics import confusion_matrix
cm = confusion_matrix(y_test, y_pred)
# Visualising the Training set results
from matplotlib.colors import ListedColormap
X_set, y_set = X_train, y_train
X1, X2 = np.meshgrid(np.arange(start = X_set[:, 0].min() - 1, stop = X_set[:, 0].max() + 1, step = 0.01),
np.arange(start = X_set[:, 1].min() - 1, stop = X_set[:, 1].max() + 1, step = 0.01))
plt.contourf(X1, X2, np.array(classifier.predict(np.array([X1.ravel(), X2.ravel()]).T)).reshape(X1.shape),
alpha = 0.75, cmap = ListedColormap(('red', 'green')))
plt.xlim(X1.min(), X1.max())
plt.ylim(X2.min(), X2.max())
for i, j in enumerate(np.unique(y_set)):
import numpy as np import matplotlib.pyplot as plt # import seaborn as sns from sklearn import datasets from Logistic_Regression import LogisticRegression iris = datasets.load_iris() X = iris.data[:, :2] y = (iris.target != 0) *1 clf = LogisticRegression() clf.fit(X,y) pred = clf.predict(X) plt.figure(figsize=(10, 6)) plt.scatter(X[y == 0][:, 0], X[y == 0][:, 1], color='b', label='0') plt.scatter(X[y == 1][:, 0], X[y == 1][:, 1], color='r', label='1') plt.legend() x1_min, x1_max = X[:,0].min(), X[:,0].max(), x2_min, x2_max = X[:,1].min(), X[:,1].max(), xx1, xx2 = np.meshgrid(np.linspace(x1_min, x1_max), np.linspace(x2_min, x2_max)) grid = np.c_[xx1.ravel(), xx2.ravel()] probs = clf.predict_prob(grid).reshape(xx1.shape) plt.contour(xx1, xx2, probs, [0.5], linewidths=1, colors='black') plt.show()
def adaboost(x_ada, y_ada, x_test_in, y_test_in):
# 初始化权重
weight = np.ones((np.size(x_ada, axis=0), 1))
weight /= np.size(x_ada, axis=0)
weight_list = []
classifier_list = []
# 训练算法
clf1 = LogisticRegression().fit(x_ada, y_ada)
predict1 = clf1.predict(x_ada, y_ada)
clf2 = LinearSVM().fit(x_ada, y_ada)
predict2 = clf2.predict(x_ada, y_ada)
clf3 = CartDecisionTree().fit(x_ada, y_ada)
predict3 = clf3.predict(x_ada, y_ada)
# 组合分类器
for i in range(Adaboost_EPOCH):
e1 = 0
e2 = 0
e3 = 0
# 计算误差
for j in range(np.size(x_ada, axis=0)):
if predict1[j] != y_ada[j]:
e1 += weight[j]
if predict2[j] != y_ada[j]:
e2 += weight[j]
if predict3[j] != y_ada[j]:
e3 += weight[j]
# 选择小误差的模型
if e1[0] <= e2[0] and e1[0] <= e3[0]:
clf = clf1
a = 1 / 2 * np.log((1 - e1[0]) / e1[0])
predict = predict1
elif e2[0] <= e1[0] and e2[0] <= e3[0]:
clf = clf2
a = 1 / 2 * np.log((1 - e2[0]) / e2[0])
predict = predict2
else:
clf = clf3
a = 1 / 2 * np.log((1 - e3[0]) / e3[0])
predict = predict3
# 更新权重
z = np.sum(np.exp(-a * (y_ada - 0.5) * (predict - 0.5) * 4),
axis=0) # x, y化成０或１
weight = weight * np.exp(-a * (y_ada - 0.5) * (predict - 0.5) * 4) / z
weight_list.append(a)
classifier_list.append(clf)
# 评估acc
predict_sum = 0
predict_get = np.zeros_like(y_test_in)
acc_count = 0
for l in range(Adaboost_EPOCH):
predict_sum += weight_list[l] * (
classifier_list[l].predict(x_test_in, y_test_in) - 0.5) * 2
for k in range(np.size(y_test_in, axis=0)):
if predict_sum[k] / Adaboost_EPOCH >= 0:
predict_get[k] = 1
else:
predict_get[k] = 0
if predict_get[k] == y_test_in[k]:
acc_count += 1
acc = acc_count / np.size(y_test_in, axis=0) * 100
print('Adaboost ACC: %.2f%%' % acc)
return 0
X_train_pca = pca.fit_transform(X_train)
X_test_pca = pca.transform(X_test)
f_measure_test = []
f_measure_train = []
Lambda = []
# Training logistic regression classifier with L2 penalty
for i in float_range(-2, -0.2, 0.2):
C_ = 1 / i
LR = LogisticRegression(learningRate=0.1, numEpochs=10, penalty='L2',
C=i) # range from 0.01 - 0.03
LR.train(X_train_pca, y_train, tol=10**-3)
# LR.plotCost()
# Testing fitted model on test data with cutoff probability 50%
predictions, probs = LR.predict(X_test_pca, 0.5)
performance = LR.performanceEval(predictions, y_test)
# added
predictions_train, probs_train = LR.predict(X_train_pca, 0.5)
performance_train = LR.performanceEval(predictions_train, y_train)
# LR.plotDecisionRegions(X_test_pca, y_test)
# LR.predictionPlot(X_test_pca, y_test)
# Print out performance values
for key, value in performance.items():
print('%s : %.2f' % (key, value))
print("\n")
f_measure_test.append(list(performance.values())[len(performance) - 1])
f_measure_train.append(
list(performance_train.values())[len(performance_train) - 1])
Lambda.append(i)