sk-特征选择.py

# # ============去除方差小于阈值的特征============
# from sklearn.feature_selection import VarianceThreshold  #移除低方差特征
# from sklearn.datasets import load_iris  # 引入花卉数据集
# iris = load_iris()
# X= iris.data
# print(X.shape)
# print(X.var(axis=0))
#
# sel = VarianceThreshold(threshold=0.2)
# X_transformed=sel.fit_transform(X)
# print('去除低方差特征：\n',X_transformed.shape)





# # ============排序选择优秀特征============
# from sklearn.datasets import load_iris
# from sklearn.feature_selection import SelectKBest
# from sklearn.feature_selection import chi2  # 引入卡方检验统计量
# # 对于回归: f_regression , mutual_info_regression
# # 对于分类: chi2 , f_classif , mutual_info_classif
# iris = load_iris()
# X, y = iris.data, iris.target
# print('源样本维度：',X.shape)
#
# X_new = SelectKBest(chi2, k=2).fit_transform(X, y)
# print('新样本维度：',X_new.shape)




# # ============递归式特征消除============
# # 这里递归的移除最不重要的像素点来对每个像素点（特征）进行排序
# from sklearn.svm import SVC
# from sklearn.datasets import load_digits
# from sklearn.feature_selection import RFE
# import matplotlib.pyplot as plt
#
# digits = load_digits()  # 加载手写体数据集
# X = digits.images.reshape((len(digits.images), -1))
# y = digits.target
#
# # 创建ref对象和每个像素点的重要度排名
# svc = SVC(kernel="linear", C=1)
# rfe = RFE(estimator=svc, n_features_to_select=1, step=1)
# rfe.fit(X, y)
# ranking = rfe.ranking_.reshape(digits.images[0].shape)
#
# # 绘制像素点排名
# plt.matshow(ranking, cmap=plt.cm.Blues)
# plt.colorbar()
# plt.title("Ranking of pixels with RFE")
# plt.show()





# # ============使用 SelectFromModel 选取特征============
#
# import matplotlib.pyplot as plt
# from sklearn.datasets import load_boston
# from sklearn.feature_selection import SelectFromModel
# from sklearn.linear_model import LassoCV
# boston = load_boston()  # 加载波士顿房价回归数据
# X, y = boston['data'], boston['target']  # 取特征数据和输出数据
# n_features =[13]  # 记录循环中的特征个数，最开始数据集是有13个特征的
# thresholds=[0]  # 记录门限值，最开始是没有门限值的
#
# clf = LassoCV()  # 使用Lasso回归
#
# # 设置最小门限为0.25。coef_ 或者 featureimportances 属性值低于门限的都会被去除调
# sfm = SelectFromModel(clf, threshold=0.1)
# sfm.fit(X, y)  # 训练模型。找出模型回归系数。
# X_transform = sfm.transform(X) # 根据回归系数、门限，变换数据集。
# n_feature =X_transform.shape[1]  # 获取训练以后的特征数目
# n_features.append(n_feature)
# thresholds.append(0.1)
# while n_feature > 2:  # 如果特征数大于2，则从新转换，找最好的两个特征
#     sfm.threshold += 0.1  # 逐渐增加门限，进一步减少特征数目
#     X_transform = sfm.transform(X) # 变换数据集
#     n_feature = X_transform.shape[1]
#     n_features.append(n_feature)  # 记录训练以后的特征数目
#     thresholds.append(sfm.threshold)  # 记录门限值
#
# plt.title("Features with threshold %0.3f." % sfm.threshold)
# plt.plot(thresholds, n_features, 'r')
# plt.xlabel("thresholds")
# plt.ylabel("Feature number")
# plt.show()




# # ============基于 L1 的特征选取============
# from sklearn.svm import LinearSVC
# from sklearn.datasets import load_iris
# from sklearn.feature_selection import SelectFromModel
# iris = load_iris()
# X, y = iris.data, iris.target
# print('原数据集维度：',X.shape)
# lsvc = LinearSVC(C=0.01, penalty="l1", dual=False).fit(X, y)
# model = SelectFromModel(lsvc, prefit=True)
# X_new = model.transform(X)
# print('新数据集维度：',X_new.shape)



# ============基于 Tree（树）的特征选取============
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.datasets import load_iris
from sklearn.feature_selection import SelectFromModel
dataset = load_iris()
X, y = dataset.data, dataset.target
print('原数据集维度：',X.shape)
clf = ExtraTreesClassifier()
clf = clf.fit(X, y)
print('属性重要程度：',clf.feature_importances_)

model = SelectFromModel(clf, prefit=True)
X_new = model.transform(X)
print('新数据集维度：',X.shape)



# ============特征选取作为 pipeline（管道）的一部分============
# from sklearn.pipeline import Pipeline
# clf = Pipeline([
#   ('feature_selection', SelectFromModel(LinearSVC(penalty="l1"))),
#   ('classification', RandomForestClassifier())
# ])
# clf.fit(X, y)