#%% 基礎(chǔ)數(shù)組運算庫導(dǎo)入
import numpy as np 
# 畫圖庫導(dǎo)入
import matplotlib.pyplot as plt 
# 導(dǎo)入三維顯示工具
from mpl_toolkits.mplot3d import Axes3D
# 導(dǎo)入BP模型
from sklearn.neural_network import MLPClassifier
# 導(dǎo)入demo數(shù)據(jù)制作方法
from sklearn.datasets import make_classification
from sklearn.metrics import classification_report, confusion_matrix
import seaborn as sns
import warnings
from sklearn.exceptions import ConvergenceWarning

#%%模型訓(xùn)練
# 制作五個類別的數(shù)據(jù)，每個類別1000個樣本
train_samples, train_labels = make_classification(n_samples=1000, n_features=3, 
                           n_redundant=0,n_classes=5, n_informative=3, 
                           n_clusters_per_class=1,class_sep=3, random_state=10)
# 將五個類別的數(shù)據(jù)進行三維顯示
fig = plt.figure()
ax = Axes3D(fig, rect=[0, 0, 1, 1], elev=20, azim=20)
ax.scatter(train_samples[:, 0], train_samples[:, 1], train_samples[:, 2], marker='o', c=train_labels)
plt.title('Demo Data Map')

#%% 建立 BP 模型, 采用sgd優(yōu)化器，relu非線性映射函數(shù)
BP = MLPClassifier(solver='sgd',activation = 'relu',max_iter = 500,alpha = 1e-3,
                   hidden_layer_sizes = (32,32),random_state = 1)
# 進行模型訓(xùn)練
with warnings.catch_warnings():
    warnings.filterwarnings("ignore", category=ConvergenceWarning,
                            module="sklearn")
    BP.fit(train_samples, train_labels)
# 查看 BP 模型的參數(shù)
print(BP)
#%% 進行模型預(yù)測
predict_labels = BP.predict(train_samples)
# 顯示預(yù)測的散點圖
fig = plt.figure()
ax = Axes3D(fig, rect=[0, 0, 1, 1], elev=20, azim=20)
ax.scatter(train_samples[:, 0], train_samples[:, 1], train_samples[:, 2], marker='o', c=predict_labels)
plt.title('Demo Data Predict Map with BP Model')

# 顯示預(yù)測分數(shù)
print("預(yù)測準確率: {:.4f}".format(BP.score(train_samples, train_labels)))

# 可視化預(yù)測數(shù)據(jù) 
print("真實類別：", train_labels[:10])
print("預(yù)測類別：", predict_labels[:10])
# 準確率等報表
print(classification_report(train_labels, predict_labels))

# 計算混淆矩陣
classes = [0, 1, 2, 3]
cofusion_mat = confusion_matrix(train_labels, predict_labels, classes) 
sns.set()
figur, ax = plt.subplots()
# 畫熱力圖
sns.heatmap(cofusion_mat, cmap="YlGnBu_r", annot=True, ax=ax) 
ax.set_title('confusion matrix')  # 標題
ax.set_xticklabels([''] + classes, minor=True)
ax.set_yticklabels([''] + classes, minor=True)
ax.set_xlabel('predict')  # x軸
ax.set_ylabel('true')  # y軸
plt.show()

#%%# 進行新的測試數(shù)據(jù)測試
test_sample = np.array([[-1, 0.1, 0.1]])
print(f"{test_sample} 類別是: ", BP.predict(test_sample))
print(f"{test_sample} 類別概率分別是: ", BP.predict_proba(test_sample))

test_sample = np.array([[-1.2, 10, -91]])
print(f"{test_sample} 類別是: ", BP.predict(test_sample))
print(f"{test_sample} 類別概率分別是: ", BP.predict_proba(test_sample))

test_sample = np.array([[-12, -0.1, -0.1]])
print(f"{test_sample} 類別是: ", BP.predict(test_sample))
print(f"{test_sample} 類別概率分別是: ", BP.predict_proba(test_sample))

test_sample = np.array([[100, -90.1, -9.1]])
print(f"{test_sample} 類別是: ", BP.predict(test_sample))
print(f"{test_sample} 類別概率分別是: ", BP.predict_proba(test_sample))

#%%基于BP神經(jīng)網(wǎng)絡(luò)的乳腺癌分類
#基本庫導(dǎo)入
# 導(dǎo)入乳腺癌數(shù)據(jù)集
from sklearn.datasets import load_breast_cancer
# 導(dǎo)入BP模型
from sklearn.neural_network import MLPClassifier
# 導(dǎo)入訓(xùn)練集分割方法
from sklearn.model_selection import train_test_split 
# 導(dǎo)入預(yù)測指標計算函數(shù)和混淆矩陣計算函數(shù)
from sklearn.metrics import classification_report, confusion_matrix
# 導(dǎo)入繪圖包
import seaborn as sns
import matplotlib.pyplot as plt
# 導(dǎo)入三維顯示工具
from mpl_toolkits.mplot3d import Axes3D
# 導(dǎo)入乳腺癌數(shù)據(jù)集
cancer = load_breast_cancer()
# 查看數(shù)據(jù)集信息
print('breast_cancer數(shù)據(jù)集的長度為：',len(cancer))
print('breast_cancer數(shù)據(jù)集的類型為：',type(cancer))
# 分割數(shù)據(jù)為訓(xùn)練集和測試集
cancer_data = cancer['data']
print('cancer_data數(shù)據(jù)維度為：',cancer_data.shape)
cancer_target = cancer['target']
print('cancer_target標簽維度為：',cancer_target.shape)
cancer_names = cancer['feature_names']
cancer_desc = cancer['DESCR']
#分為訓(xùn)練集與測試集
cancer_data_train,cancer_data_test = train_test_split(cancer_data,test_size=0.2,random_state=42)#訓(xùn)練集
cancer_target_train,cancer_target_test = train_test_split(cancer_target,test_size=0.2,random_state=42)#測試集

#%%# 建立 BP 模型, 采用Adam優(yōu)化器，relu非線性映射函數(shù)
BP = MLPClassifier(solver='adam',activation = 'relu',max_iter = 1000,alpha = 1e-3,hidden_layer_sizes = (64,32, 32),random_state = 1)
# 進行模型訓(xùn)練
BP.fit(cancer_data_train, cancer_target_train)
#%% 進行模型預(yù)測
predict_train_labels = BP.predict(cancer_data_train)
# 可視化真實數(shù)據(jù)
fig = plt.figure()
ax = Axes3D(fig, rect=[0, 0, 1, 1], elev=20, azim=20) 
ax.scatter(cancer_data_train[:, 0], cancer_data_train[:, 1], cancer_data_train[:, 2], marker='o', c=cancer_target_train)
plt.title('True Label Map')
plt.show()
# 可視化預(yù)測數(shù)據(jù)
fig = plt.figure()
ax = Axes3D(fig, rect=[0, 0, 1, 1], elev=20, azim=20) 
ax.scatter(cancer_data_train[:, 0], cancer_data_train[:, 1], cancer_data_train[:, 2], marker='o', c=predict_train_labels)
plt.title('Cancer with BP Model')
plt.show()

#%% 顯示預(yù)測分數(shù)
print("預(yù)測準確率: {:.4f}".format(BP.score(cancer_data_test, cancer_target_test)))
# 進行測試集數(shù)據(jù)的類別預(yù)測
predict_test_labels = BP.predict(cancer_data_test)
print("測試集的真實標簽:\n", cancer_target_test)
print("測試集的預(yù)測標簽:\n", predict_test_labels)
#%% 進行預(yù)測結(jié)果指標統(tǒng)計 統(tǒng)計每一類別的預(yù)測準確率、召回率、F1分數(shù)
print(classification_report(cancer_target_test, predict_test_labels))

#%% 計算混淆矩陣
confusion_mat = confusion_matrix(cancer_target_test, predict_test_labels)
# 打印混淆矩陣
print(confusion_mat)
# 將混淆矩陣以熱力圖的方式顯示
sns.set()
figure, ax = plt.subplots()
# 畫熱力圖
sns.heatmap(confusion_mat, cmap="YlGnBu_r", annot=True, ax=ax)  
# 標題 
ax.set_title('confusion matrix')
# x軸為預(yù)測類別
ax.set_xlabel('predict')  
# y軸實際類別
ax.set_ylabel('true')  
plt.show()