python数据分析笔记——第十章 预测性分析和机器学习

2.预处理

import numpy as np
from sklearn import preprocessing
from scipy.stats import anderson
# 加载数据
rain = np.load('rain.npy')
rain = .1 * rain
rain[rain < 0] = .05 / 2
# 期望值 标准差和安德森
print("Rain mean", rain.mean())
print("Rain Variance", rain.var())
print("Anderson Rain", anderson(rain))
#对数据缩放处理
scaled = preprocessing.scale(rain)
print("Scaled mean", scaled.mean())
print("Scaled Variance", scaled.var())
print("Anderson Scaled", anderson(scaled))
# 把特征值从数值型转换布尔型
binarized = preprocessing.binarize(rain)
print("binarized", np.unique(binarized), binarized.sum())
# 用整数来标注类别
lb = preprocessing.LabelBinarizer()
lb.fit(rain.astype(int))
print(lb.classes_)

3.基于逻辑回归的分类
该算法可以用以预测事件发生的概率,或是事物是否属于一类别的概率

from sklearn.linear_model import LogisticRegression
from sklearn.cross_validation import KFold
from sklearn import datasets
import numpy as np

def classify(x, y):
#使用逻辑回归进行分类
    clf = LogisticRegression(random_state=12)
    scores = []
# k-折交叉验证
    kf = KFold(len(y), n_folds=10)
# 检查分类的状确性
    for train, test in kf:
        clf.fit(x[train], y[train])
        scores.append(clf.score(x[test], y[test]))
    print(np.mean(scores))

#加载数据信息
rain = np.load('rain.npy')
dates = np.load('doy.npy')
#使用日期和降雨量来构建数组
x = np.vstack((dates[:-1], rain[:-1]))
# 无雨，小雨，雨
y = np.sign(rain[1:])
classify(x.T, y)
#sklearn数据集
iris = datasets.load_iris()
x = iris.data[:, :2]
y = iris.target
classify(x, y)

4.基于支持向量机的分类
支持向量机 Support vector machines SVM
支持向量回归 Support vector Regression SVR
可以用来进行回归分析,也可以用来分类

from sklearn.svm import SVC
from sklearn.grid_search import GridSearchCV
from sklearn import datasets
import numpy as np
from pprint import PrettyPrinter

def classify(x, y):
    # 进行网格搜索
    clf = GridSearchCV(SVC(random_state=42, max_iter=100), {'kernel': ['linear', 'poly', 'rbf'], 'C': [1, 10]})
    clf.fit(x, y)
    print("Score", clf.score(x, y))
    PrettyPrinter().pprint(clf.grid_scores_)

rain = np.load('rain.npy')
dates = np.load('doy.npy')
x = np.vstack((dates[:-1], rain[:-1]))
y = np.sign(rain[1:])
classify(x.T, y)

iris = datasets.load_iris()
x = iris.data[:, :2]
y = iris.target
classify(x, y)

5.基于elasticNetCV的回归分类
弹性网格正则化 Elasic net Regularization 降低回归分析的过拟合风险
实际上是LASSO(The Least Absolute Shrikage and Selection Operator)算法和岭回归方法的线性组合。

from sklearn.linear_model import ElasticNetCV
import numpy as np
from sklearn import datasets
import matplotlib.pyplot as plt

def regress(x, y, title):
    clf = ElasticNetCV(max_iter=200,# 最大迭代次数
                       cv=10,# 包总量
                       l1_ratio=[.1, .5, .7, .9, .95, .99, 1])
                       # 0表示只使用岭回归,1表示只使用 LASSO回归,否则使用混合算法
    clf.fit(x, y)
    print("Score", clf.score(x, y))

    pred = clf.predict(x)
    plt.title("Scatter plot of prediction and " + title)
    plt.xlabel("Prediction")
    plt.ylabel("Target")
    plt.scatter(y, pred)

    if "Boston" in title:
        plt.plot(y, y, label="Perfect Fit")
        plt.legend()
    plt.grid = True
    plt.show()

rain = .1 * np.load('rain.npy')
rain[rain < 0] = .05 / 2
dates = np.load("doy.npy")

x = np.vstack((dates[:-1], rain[:-1]))
y = rain[1:]
regress(x.T, y, "rain data")

boston = datasets.load_boston()
x = boston.data
y = boston.target
regress(x, y, "Boston house prices")

6.支持向量回归

import numpy as np
from sklearn import datasets
from sklearn.learning_curve import learning_curve
from sklearn.svm import SVR
from sklearn import preprocessing
import multiprocessing
import matplotlib.pyplot as plt

def regress(x, y, ncpus, title):
    X = preprocessing.scale(x)
    Y = preprocessing.scale(y)
    clf = SVR(max_iter=ncpus * 200)
    # 根据cpu数量创建作业数
    train_sizes, train_scores, test_scores = learning_curve(clf, X, Y, n_jobs=ncpus)

    # 求平均数,然后画出得分
    plt.figure()
    plt.title(title)
    plt.plot(train_sizes, train_scores.mean(axis=1), label="Train score")
    plt.plot(train_sizes, test_scores.mean(axis=1), '--', label="Test score")
    print("Max test score " + title, test_scores.max())
    plt.legend(loc='best')
    plt.show()

def main():
    rain = .1 * np.load('rain.npy')
    rain[rain < 0] = .05 / 2
    dates = np.load('doy.npy')

    x = np.vstack((dates[:-1], rain[:-1]))
    y = rain[1:]
    ncpus = multiprocessing.cpu_count()
    regress(x.T, y, ncpus, "Rain")

    boston = datasets.load_boston()
    x = boston.data
    y = boston.target
    regress(x, y, ncpus, "Boston")

if __name__ == '__main__':
    main()

7.基于相似性传播算法的聚类分析
聚类分析就是把数据分成一些组,这些组就是所谓的聚类
聚类分析，属无监督学习
相似性传播 affinity propagation

import numpy as np
from sklearn import datasets
from sklearn import cluster
from sklearn.metrics import euclidean_distances
import matplotlib.pyplot as plt

# 生成三个数据块
x, _ = datasets.make_blobs(n_samples=100, centers=3, n_features=2, random_state=10)
# 创建矩阵
S = euclidean_distances(x)
# print(S)

# 根据矩阵,给数据标注其所属聚类
aff_pro = cluster.AffinityPropagation().fit(S)
labels = aff_pro.labels_

# 绘制图形
styles = ['o', 'x', '^']
for style, label in zip(styles, np.unique(labels)):
    print(label)
    plt.plot(x[labels == label], style, label=label)

plt.title("Clustering Blobs")
plt.legend(loc='best')
plt.show()

8 均值漂移算法
一种不需要估算聚类数的聚类算法。

import numpy as np
from sklearn import cluster
import matplotlib.pyplot as plt
import pandas as pd

# 加载数据
rain = .1 * np.load('rain.npy')
rain[rain < 0] = .05 / 2
dates = np.load('doy.npy')
x = np.vstack((dates, rain))

# 创建dataFrame,并计算平均值
df = pd.DataFrame.from_records(x.T, columns=['dates', 'rain'])
df = df.groupby('dates').mean()
df.plot()

# 均值漂移算法
x = np.vstack((np.arange(1, len(df) + 1), df.as_matrix().ravel()))
x = x.T
ms = cluster.MeanShift()
ms.fit(x)
labels = ms.predict(x)

# 绘制图形
plt.figure()
grays = ['0', '0.5', '0.75']

for gray, label in zip(grays, np.unique(labels)):
    match = labels == label
    x0 = x[:, 0]
    x1 = x[:, 1]
    plt.plot(x0[match], x1[match], lw=label + 1, label=label)
    plt.fill_between(x0, x1, where=match, color=gray)

plt.legend()
plt.show()

9.遗传算法

import array
import random
import numpy as np
from deap import algorithms
from deap import base
from deap import creator
from deap import tools
from scipy.stats import shapiro
import matplotlib.pyplot as plt


creator.create("FitnessMax", base.Fitness, weights=(1.0,))
creator.create("Individual", array.array, typecode='d', fitness=creator.FitnessMax)

toolbox = base.Toolbox()
toolbox.register("attr_float", random.random)
toolbox.register("individual", tools.initRepeat, creator.Individual, toolbox.attr_float, 200)
toolbox.register("populate", tools.initRepeat, list, toolbox.individual)

def eval(individual):
    return shapiro(individual)[1],

toolbox.register("evaluate", eval)
toolbox.register("mate", tools.cxTwoPoint)
toolbox.register("mutate", tools.mutFlipBit, indpb=0.1)
toolbox.register("select", tools.selTournament, tournsize=4)

random.seed(42)

pop = toolbox.populate(n=400)
hof = tools.HallOfFame(1)
stats = tools.Statistics(key=lambda ind: ind.fitness.values)
stats.register("max", np.max)

algorithms.eaSimple(pop, toolbox, cxpb=0.5, mutpb=0.2, ngen=80, stats=stats, halloffame=hof)

print (shapiro(hof[0])[1])
plt.hist(hof[0])
plt.grid(True)
plt.show()

10.神经网络

import numpy as np
import theanets
import multiprocessing
from sklearn import datasets
from sklearn.metrics import accuracy_score


rain = .1 * np.load('rain.npy')
rain[rain < 0] = .05/2
dates = np.load('doy.npy')
x = np.vstack((dates[:-1], np.sign(rain[:-1])))
x = x.T

y = np.vstack(np.sign(rain[1:]),)
N = int(.9 * len(x))

e = theanets.Experiment(theanets.Regressor,
                        layers=(2, 3, 1),
                        learning_rate=0.1,
                        momentum=0.5,
                        patience=300,
                        train_batches=multiprocessing.cpu_count(),
                        num_updates=500)

train = [x[:N], y[:N]]
valid = [x[N:], y[N:]]
e.run(train, valid)

pred = e.network(x[N:]).ravel()
print ("Pred Min", pred.min(), "Max", pred.max())
print ("Y Min", y.min(), "Max", y.max())
print ("Accuracy", accuracy_score(y[N:], pred >= .5))

11.决策树

#所属模块发生变化
from sklearn.cross_validation import train_test_split
from sklearn.grid_search import RandomizedSearchCV
from sklearn import tree
from scipy.stats import randint as sp_randint
import pydot
# import StringIO
from io import StringIO
import numpy as np
from tempfile import NamedTemporaryFile

# 加载数据信息
rain = .1 * np.load('rain.npy')
rain[rain < 0] = .05 / 2

dates = np.load('doy.npy').astype(int)
x = np.vstack((dates[:-1], np.sign(rain[:-1])))
x = x.T

y = np.sign(rain[1:])

# 创建测试集和训练集数据
x_tain, x_test, y_train, y_test = train_test_split(x, y, random_state=37)

# 验证各参数的取值范围
clf = tree.DecisionTreeClassifier(random_state=37)
params = {"max_depth": [2, None],"min_samples_leaf": sp_randint(1, 5),"criterion": ["gini", "entropy"]}
rscv = RandomizedSearchCV(clf, params)
rscv.fit(x_tain, y_train)

# 绘制决策树的对象
sio = StringIO()
tree.export_graphviz(rscv.best_estimator_, out_file=sio, feature_names=['day-of-year', 'yest'])
dec_tree = pydot.graph_from_dot_data(sio.getvalue())
with NamedTemporaryFile(prefix='rain', suffix='.png', delete=False) as f:
    # dec_tree.write_png(f.name)
    dec_tree[0].write_png(f.name)
    print("Written figure to", f.name)

print('Best Train Score', rscv.best_score_)
print('Test Score', rscv.score(x_test, y_test))
print("Best params", rscv.best_params_)