Scikit-learn 是開源的 Python 庫,經過統一的界面實現機器學習、預處理、交叉驗證及可視化算法。python
scikit-learn 網站:scikit-learn.org算法
Python 中的機器學習數組
肯定對象屬於哪一個類別。dom
應用:垃圾郵件檢測,圖像識別。機器學習
算法: SVM,最近鄰居,隨機森林,......ide
預測與對象關聯的連續值屬性。函數
應用:藥物反應,股票價格。工具
算法: SVR,嶺迴歸,套索,......性能
將相似對象自動分組到集合中。學習
應用:客戶細分,分組實驗結果
算法: k-Means,譜聚類,均值漂移,......
減小要考慮的隨機變量的數量。
應用:可視化,提升效率
算法: PCA,特徵選擇,非負矩陣分解。
比較,驗證和選擇參數和模型。
目標:經過參數調整提升準確性
模塊: 網格搜索,交叉驗證,指標。
特徵提取和規範化。
應用程序:轉換輸入數據(如文本)以與機器學習算法一塊兒使用。 模塊: 預處理,特徵提取。
# 導入 sklearn
from sklearn import neighbors, datasets, preprocessing
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
# 加載數據
iris = datasets.load_iris()
# 劃分訓練集與測試集
X, y = iris.data[:, :2], iris.target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=33)
# 數據預處理
scaler = preprocessing.StandardScaler().fit(X_train)
X_train = scaler.transform(X_train)
X_test = scaler.transform(X_test)
# 建立模型
knn = neighbors.KNeighborsClassifier(n_neighbors=5)
# 模型擬合
knn.fit(X_train, y_train)
# 預測
y_pred = knn.predict(X_test)
# 評估
accuracy_score(y_test, y_pred)
複製代碼
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
複製代碼
Scikit-learn 處理的數據是存儲爲 NumPy 數組或 SciPy 稀疏矩陣的數字,還支持 Pandas 數據框等可轉換爲數字數組的其它數據類型。
X = np.random.random((11, 5))
y = np.array(['M', 'M', 'F', 'F', 'M', 'F', 'M', 'M', 'F', 'F', 'F'])
X[X < 0.7] = 0
複製代碼
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)
複製代碼
from sklearn.preprocessing import StandardScaler
scaler = StandardScaler().fit(X_train)
standardized_X = scaler.transform(X_train)
standardized_X_test = scaler.transform(X_test)
複製代碼
from sklearn.preprocessing import Normalizer
scaler = Normalizer().fit(X_train)
normalized_X = scaler.transform(X_train)
normalized_X_test = scaler.transform(X_test)
複製代碼
from sklearn.preprocessing import Binarizer
binarizer = Binarizer(threshold=0.0).fit(X)
binary_X = binarizer.transform(X)
複製代碼
from sklearn.preprocessing import LabelEncoder
enc = LabelEncoder()
y = enc.fit_transform(y)
複製代碼
from sklearn.preprocessing import Imputer
imp = Imputer(missing_values=0, strategy='mean', axis=0)
imp.fit_transform(X_train)
複製代碼
from sklearn.preprocessing import PolynomialFeatures
poly = PolynomialFeatures(5)
poly.fit_transform(X)
複製代碼
# 線性迴歸
from sklearn.linear_model import LinearRegression
lr = LinearRegression(normalize=True)
# 支持向量機(SVM)
from sklearn.svm import SVC
svc = SVC(kernel='linear')
# 樸素貝葉斯
from sklearn.naive_bayes import GaussianNB
gnb = GaussianNB()
# KNN
from sklearn import neighbors
knn = neighbors.KNeighborsClassifier(n_neighbors=5)
複製代碼
# 主成分分析(PCA)
from sklearn.cluster import KMeans
from sklearn.decomposition import PCA
pca = PCA(n_components=0.95)
# K Means
k_means = KMeans(n_clusters=3, random_state=0)
複製代碼
lr.fit(X, y)
knn.fit(X_train, y_train)
svc.fit(X_train, y_train)
複製代碼
k_means.fit(X_train)
pca_model = pca.fit_transform(X_train)
複製代碼
# 預測標籤
y_pred = svc.predict(np.random.random((2,5)))
# 預測標籤
y_pred = lr.predict(X_test)
# 評估標籤機率
y_pred = knn.predict_proba(X_test)
複製代碼
y_pred = k_means.predict(X_test)
複製代碼
# 準確率
knn.score(X_test, y_test)
from sklearn.metrics import accuracy_score
accuracy_score(y_test, y_pred)
# 分類預估評價函數
from sklearn.metrics import classification_report
print(classification_report(y_test, y_pred))
# 混淆矩陣
from sklearn.metrics import confusion_matrix
print(confusion_matrix(y_test, y_pred))
複製代碼
# 平均絕對偏差
from sklearn.metrics import mean_absolute_error
y_true = [3, -0.5, 2]
mean_absolute_error(y_true, y_pred)
# 均方偏差
from sklearn.metrics import mean_squared_error
mean_squared_error(y_test, y_pred)
# R2 評分
from sklearn.metrics import r2_score
r2_score(y_true, y_pred)
複製代碼
# 調整蘭德係數
from sklearn.metrics import adjusted_rand_score
adjusted_rand_score(y_true, y_pred)
# 同質性
from sklearn.metrics import homogeneity_score
homogeneity_score(y_true, y_pred)
# V-measure
from sklearn.metrics import v_measure_score
metrics.v_measure_score(y_true, y_pred)
複製代碼
from sklearn.cross_validation import cross_val_score
print(cross_val_score(knn, X_train, y_train, cv=4))
print(cross_val_score(lr, X, y, cv=2))
複製代碼
from sklearn.grid search import GridSearchcV
params = {"n neighbors": np.arange(1, 3),
"metric": ["euclidean", "cityblock"]}
grid = GridSearchCV(estimator=knn,
param_grid-params)
grid.fit(X_train, y_train)
print(grid.best score)
print(grid.best_estimator_.n_neighbors)
複製代碼
from sklearn.grid_search import RandomizedSearchCV
params = {"n_neighbors": range(1, 5),
"weights": ["uniform", "distance"]}
rsearch = RandomizedSearchCV(estimator=knn,
rsearch.fit(X_train, y_train) random_state=5)
print(rsearch.best_score_)
複製代碼