机器学习04-(决策树、集合算法:AdaBoost模型、BBDT、随机森林、分类模型:逻辑回归)

机器学习-04

集合算法

根据多个不同模型给出的预测结果,利用平均(回归)或者投票(分类)的方法,得出最终预测结果。

基于决策树的集合算法,就是按照某种规则,构建多棵彼此不同的决策树模型,分别给出针对未知样本的预测结果,最后通过平均或投票得到相对综合的结论。常用的集合模型包括Boosting类模型(AdaBoost、GBDT)与Bagging(自助聚合、随机森林)类模型。

AdaBoost模型(正向激励)

首先为样本矩阵中的样本随机分配初始权重,由此构建一棵带有权重的决策树,在由该决策树提供预测输出时,通过加权平均或者加权投票的方式产生预测值。将训练样本代入模型,预测其输出,对那些预测值与实际值不同的样本,提高其权重,由此形成第二棵决策树。重复以上过程,构建出不同权重的若干棵决策树。

正向激励相关API:

import sklearn.tree as st
import sklearn.ensemble as se
# model: 决策树模型(一颗)
model = st.DecisionTreeRegressor(max_depth=4)
# 自适应增强决策树回归模型	
# n_estimators:构建400棵不同权重的决策树,训练模型
model = se.AdaBoostRegressor(model, n_estimators=400, random_state=7)
# 训练模型
model.fit(train_x, train_y)
# 测试模型
pred_test_y = model.predict(test_x)

案例:基于正向激励训练预测波士顿地区房屋价格的模型。

# 创建基于决策树的正向激励回归器模型
model = se.AdaBoostRegressor(
	st.DecisionTreeRegressor(max_depth=4), n_estimators=400, random_state=7)
# 训练模型
model.fit(train_x, train_y)
# 测试模型
pred_test_y = model.predict(test_x)
print(sm.r2_score(test_y, pred_test_y))
特征重要性

作为决策树模型训练过程的副产品,根据划分子表时选择特征的顺序标志了该特征的重要程度,此即为该特征重要性指标。训练得到的模型对象提供了属性:feature_importances_来存储每个特征的重要性。

获取样本矩阵特征重要性属性:

model.fit(train_x, train_y)
fi = model.feature_importances_

案例:获取普通决策树与正向激励决策树训练的两个模型的特征重要性值,按照从大到小顺序输出绘图。

import matplotlib.pyplot as mp

model = st.DecisionTreeRegressor(max_depth=4)
model.fit(train_x, train_y)
# 决策树回归器给出的特征重要性
fi_dt = model.feature_importances_
model = se.AdaBoostRegressor(
    st.DecisionTreeRegressor(max_depth=4), n_estimators=400, random_state=7)
model.fit(train_x, train_y)
# 基于决策树的正向激励回归器给出的特征重要性
fi_ab = model.feature_importances_

mp.figure('Feature Importance', facecolor='lightgray')
mp.subplot(211)
mp.title('Decision Tree', fontsize=16)
mp.ylabel('Importance', fontsize=12)
mp.tick_params(labelsize=10)
mp.grid(axis='y', linestyle=':')
sorted_indices = fi_dt.argsort()[::-1]
pos = np.arange(sorted_indices.size)
mp.bar(pos, fi_dt[sorted_indices], facecolor='deepskyblue', edgecolor='steelblue')
mp.xticks(pos, feature_names[sorted_indices], rotation=30)
mp.subplot(212)
mp.title('AdaBoost Decision Tree', fontsize=16)
mp.ylabel('Importance', fontsize=12)
mp.tick_params(labelsize=10)
mp.grid(axis='y', linestyle=':')
sorted_indices = fi_ab.argsort()[::-1]
pos = np.arange(sorted_indices.size)
mp.bar(pos, fi_ab[sorted_indices], facecolor='lightcoral', edgecolor='indianred')
mp.xticks(pos, feature_names[sorted_indices], rotation=30)
mp.tight_layout()
mp.show()
GBDT

GBDT(Gradient Boosting Decision Tree 梯度提升树)通过多轮迭代,每轮迭代产生一个弱分类器,每个分类器在上一轮分类器的残差**(残差在数理统计中是指实际观察值与估计值(拟合值)之间的差)**基础上进行训练。基于预测结果的残差设计损失函数。GBDT训练的过程即是求该损失函数最小值的过程。

案例:预测年龄

样本消费金额上网时长年龄
A1000114
B8001.216
C12000.924
D14001.526

训练第一颗决策树:

总 样 本 { ( 金 额 < = 1000 ) ⇒ [ A 1000 1 14 B 800 1.2 16 ] ( 金 额 > 1000 ) ⇒ [ C 1200 1 24 D 1400 1.5 26 ] 总样本\left\{ \begin{aligned} (金额<=1000) \quad \Rightarrow \quad \begin{bmatrix} A & 1000 & 1 & 14\\ B & 800 & 1.2 & 16\\ \end{bmatrix}\\ \\ (金额>1000) \quad \Rightarrow \quad \begin{bmatrix} C & 1200 & 1 & 24\\ D & 1400 & 1.5 & 26\\ \end{bmatrix}\\ \end{aligned} \right. (<=1000)[AB100080011.21416](>1000)[CD1200140011.52426]

计算每个样本的真实结果与预测结果之差(残差):

样本消费金额上网时长残差
A10001-1
B8001.21
C12000.9-1
D14001.51

基于残差训练第二颗决策树:
总 样 本 { ( 时 长 < = 1 ) ⇒ [ A 1000 1 − 1 C 1200 1 − 1 ] ( 时 长 > 1 ) ⇒ [ B 800 1.2 1 D 1400 1.5 1 ] 总样本\left\{ \begin{aligned} (时长<=1) \quad \Rightarrow \quad \begin{bmatrix} A & 1000 & 1 & -1\\ C & 1200 & 1 & -1\\ \end{bmatrix}\\ \\ (时长>1) \quad \Rightarrow \quad \begin{bmatrix} B & 800 & 1.2 & 1\\ D & 1400 & 1.5 & 1\\ \end{bmatrix}\\ \end{aligned} \right. (<=1)[AC100012001111](>1)[BD80014001.21.511]

再次计算每个样本残差为0,再无优化空间,迭代结束。

预测过程:预测以下样本的年龄:

样本消费金额上网时长年龄
E12001.6?

将每颗决策树的预测结果相加: 25 + 1 = 26 25 + 1 = 26 25+1=26

import sklearn.tree as st
import sklearn.ensemble as se
# 自适应增强决策树回归模型	
# n_estimators:构建400棵不同权重的决策树,训练模型
model = se.GridientBoostingRegressor(
    	max_depth=10, n_estimators=1000, min_samples_split=2)
# 训练模型
model.fit(train_x, train_y)
# 测试模型
pred_test_y = model.predict(test_x)
自助聚合

每次从总样本矩阵中以有放回抽样的方式随机抽取部分样本构建决策树,这样形成多棵包含不同训练样本的决策树,以削弱某些强势样本对模型预测结果的影响,提高模型的泛化特性。

随机森林

在自助聚合的基础上,每次构建决策树模型时,不仅随机选择部分样本,而且还随机选择部分特征,这样的集合算法,不仅规避了强势样本对预测结果的影响,而且也削弱了强势特征的影响,使模型的预测能力更加泛化。

随机森林相关API:

import sklearn.ensemble as se
# 随机森林回归模型	(属于集合算法的一种)
# max_depth:决策树最大深度10
# n_estimators:构建1000棵决策树,训练模型
# min_samples_split: 子表中最小样本数 若小于这个数字,则不再继续向下拆分
model = se.RandomForestRegressor(
    max_depth=10, n_estimators=1000, min_samples_split=2)

案例:分析共享单车的需求,从而判断如何进行共享单车的投放。

import numpy as np
import sklearn.utils as su
import sklearn.ensemble as se
import sklearn.metrics as sm
import matplotlib.pyplot as mp

data = np.loadtxt('../data/bike_day.csv', unpack=False, dtype='U20', delimiter=',')
day_headers = data[0, 2:13]
x = np.array(data[1:, 2:13], dtype=float)
y = np.array(data[1:, -1], dtype=float)

x, y = su.shuffle(x, y, random_state=7)
print(x.shape, y.shape)
train_size = int(len(x) * 0.9)
train_x, test_x, train_y, test_y = \
    x[:train_size], x[train_size:], y[:train_size], y[train_size:]
# 随机森林回归器
model = se.RandomForestRegressor( max_depth=10, n_estimators=1000, min_samples_split=2)
model.fit(train_x, train_y)
# 基于“天”数据集的特征重要性
fi_dy = model.feature_importances_
pred_test_y = model.predict(test_x)
print(sm.r2_score(test_y, pred_test_y))

data = np.loadtxt('../data/bike_hour.csv', unpack=False, dtype='U20', delimiter=',')
hour_headers = data[0, 2:13]
x = np.array(data[1:, 2:13], dtype=float)
y = np.array(data[1:, -1], dtype=float)
x, y = su.shuffle(x, y, random_state=7)
train_size = int(len(x) * 0.9)
train_x, test_x, train_y, test_y = \
    x[:train_size], x[train_size:], \
    y[:train_size], y[train_size:]
# 随机森林回归器
model = se.RandomForestRegressor(
    max_depth=10, n_estimators=1000,
    min_samples_split=2)
model.fit(train_x, train_y)
# 基于“小时”数据集的特征重要性
fi_hr = model.feature_importances_
pred_test_y = model.predict(test_x)
print(sm.r2_score(test_y, pred_test_y))

画图显示两组样本数据的特征重要性:

mp.figure('Bike', facecolor='lightgray')
mp.subplot(211)
mp.title('Day', fontsize=16)
mp.ylabel('Importance', fontsize=12)
mp.tick_params(labelsize=10)
mp.grid(axis='y', linestyle=':')
sorted_indices = fi_dy.argsort()[::-1]
pos = np.arange(sorted_indices.size)
mp.bar(pos, fi_dy[sorted_indices], facecolor='deepskyblue', edgecolor='steelblue')
mp.xticks(pos, day_headers[sorted_indices], rotation=30)

mp.subplot(212)
mp.title('Hour', fontsize=16)
mp.ylabel('Importance', fontsize=12)
mp.tick_params(labelsize=10)
mp.grid(axis='y', linestyle=':')
sorted_indices = fi_hr.argsort()[::-1]
pos = np.arange(sorted_indices.size)
mp.bar(pos, fi_hr[sorted_indices], facecolor='lightcoral', edgecolor='indianred')
mp.xticks(pos, hour_headers[sorted_indices], rotation=30)
mp.tight_layout()
mp.show()

分类模型

什么问题属于分类问题?

sklearn.datasets.load_iris() 鸢尾花数据集

逻辑回归

逻辑回归分类模型是一种基于回归思想实现分类业务的分类模型。

逻辑回归做二元分类时的核心思想为:

针对输出为{0, 1}的已知训练样本训练一个回归模型,使得训练样本的预测输出限制在(0, 1)的数值区间。该使原类别为0的样本的输出更接近于0,原类别为1的样本的输出更接近于1。这样就可以使用相同的回归模型来完成分类预测。

逻辑回归原理:

逻辑回归目标函数:
逻 辑 函 数 ( s i g m o i d ) : y = 1 1 + e − z ; z = w T x + b 逻辑函数(sigmoid):y = \frac{1}{1+e^{-z}}; \quad z=w^Tx+b (sigmoid)y=1+ez1;z=wTx+b
该逻辑函数值域被限制在(0, 1)区间,当x>0,y>0.5;当x<0, y<0.5。可以把训练样本数据通过线性预测模型 z z z 代入逻辑函数,找到一组最优秀的模型参数使得原本属于1类别的样本输出趋近于1;原本属于0类别的样本输出趋近于0。即将预测函数的输出看做被划分为1类的概率,择概率大的类别作为预测结果。

逻辑回归相关API:

import sklearn.linear_model as lm
# 构建逻辑回归器 
# solver:逻辑函数中指数的函数关系(liblinear为线型函数关系)
# C:参数代表正则强度,为了防止过拟合。正则越大拟合效果越小。
model = lm.LogisticRegression(solver='liblinear', C=正则强度)
model.fit(训练输入集,训练输出集)
result = model.predict(带预测输入集)

案例:基于逻辑回归器绘制网格化坐标颜色矩阵。

import numpy as np
import sklearn.linear_model as lm
import matplotlib.pyplot as mp
x = np.array([
    [3, 1],
    [2, 5],
    [1, 8],
    [6, 4],
    [5, 2],
    [3, 5],
    [4, 7],
    [4, -1]])
y = np.array([0, 1, 1, 0, 0, 1, 1, 0])`
# 逻辑分类器
model = lm.LogisticRegression(solver='liblinear', C=1)
model.fit(x, y)
l, r = x[:, 0].min() - 1, x[:, 0].max() + 1
b, t = x[:, 1].min() - 1, x[:, 1].max() + 1
n = 500
grid_x, grid_y = np.meshgrid(np.linspace(l, r, n), np.linspace(b, t, n))
samples = np.column_stack((grid_x.ravel(), grid_y.ravel()))

grid_z = model.predict(samples)
grid_z = grid_z.reshape(grid_x.shape)
mp.figure('Logistic Classification', facecolor='lightgray')
mp.title('Logistic Classification', fontsize=20)
mp.xlabel('x', fontsize=14)
mp.ylabel('y', fontsize=14)
mp.tick_params(labelsize=10)
mp.pcolormesh(grid_x, grid_y, grid_z, cmap='gray')
mp.scatter(x[:, 0], x[:, 1], c=y, cmap='brg', s=80)
mp.show()

多元分类

通过多个二元分类器解决多元分类问题。

特征1特征2==>所属类别
47==>A
3.58==>A
1.21.9==>B
5.42.2==>C

若拿到一组新的样本,可以基于二元逻辑分类训练出一个模型判断属于A类别的概率。再使用同样的方法训练出两个模型分别判断属于B、C类型的概率,最终选择概率最高的类别作为新样本的分类结果。

案例:基于逻辑分类模型的多元分类。

import numpy as np
import sklearn.linear_model as lm
import matplotlib.pyplot as mp
x = np.array([
    [4, 7],
    [3.5, 8],
    [3.1, 6.2],
    [0.5, 1],
    [1, 2],
    [1.2, 1.9],
    [6, 2],
    [5.7, 1.5],
    [5.4, 2.2]])
y = np.array([0, 0, 0, 1, 1, 1, 2, 2, 2])
# 逻辑分类器
model = lm.LogisticRegression(solver='liblinear', C=1000)
model.fit(x, y)
l, r = x[:, 0].min() - 1, x[:, 0].max() + 1
b, t = x[:, 1].min() - 1, x[:, 1].max() + 1
n = 500
grid_x, grid_y = np.meshgrid(np.linspace(l, r, n), np.linspace(b, t, n))
samples = np.column_stack((grid_x.ravel(), grid_y.ravel()))
grid_z = model.predict(samples)
print(grid_z)
grid_z = grid_z.reshape(grid_x.shape)

mp.figure('Logistic Classification', facecolor='lightgray')
mp.title('Logistic Classification', fontsize=20)
mp.xlabel('x', fontsize=14)
mp.ylabel('y', fontsize=14)
mp.tick_params(labelsize=10)
mp.pcolormesh(grid_x, grid_y, grid_z, cmap='gray')
mp.scatter(x[:, 0], x[:, 1], c=y, cmap='brg', s=80)
mp.show()

代码总结

波士顿房屋价格数据分析与房价预测

import numpy as np
import matplotlib.pyplot as plt
import sklearn.datasets as sd
import sklearn.utils as su
import pandas as pd
# 加载数据集
boston = sd.load_boston()
# print(boston.DESCR)
x, y, header = boston.data, boston.target, boston.feature_names
# 针对当前数据集,做简单的数据分析
data = pd.DataFrame(x, columns=header)
data['y'] = y
data.describe()
CRIMZNINDUSCHASNOXRMAGEDISRADTAXPTRATIOBLSTATy
count506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000506.000000
mean3.59376111.36363611.1367790.0691700.5546956.28463468.5749013.7950439.549407408.23715418.455534356.67403212.65306322.532806
std8.59678323.3224536.8603530.2539940.1158780.70261728.1488612.1057108.707259168.5371162.16494691.2948647.1410629.197104
min0.0063200.0000000.4600000.0000000.3850003.5610002.9000001.1296001.000000187.00000012.6000000.3200001.7300005.000000
25%0.0820450.0000005.1900000.0000000.4490005.88550045.0250002.1001754.000000279.00000017.400000375.3775006.95000017.025000
50%0.2565100.0000009.6900000.0000000.5380006.20850077.5000003.2074505.000000330.00000019.050000391.44000011.36000021.200000
75%3.64742312.50000018.1000000.0000000.6240006.62350094.0750005.18842524.000000666.00000020.200000396.22500016.95500025.000000
max88.976200100.00000027.7400001.0000000.8710008.780000100.00000012.12650024.000000711.00000022.000000396.90000037.97000050.000000
data.pivot_table(index='CHAS', values='y')
y
CHAS
0.022.093843
1.028.440000
data['DIS'].plot.box()
<matplotlib.axes._subplots.AxesSubplot at 0x2d7838dc5c0>

在这里插入图片描述

# 判断每个字段与房价之间的关系
data.plot.scatter(x='RM', y='y')
<matplotlib.axes._subplots.AxesSubplot at 0x2d78388b240>

在这里插入图片描述

训练回归模型,预测房屋价格

  1. 整理数据集(输入集、输出集)
  2. 打乱数据集,拆分测试集、训练集。
  3. 选择模型,使用训练集训练模型,用测试集测试。
# 整理数据集(输入集、输出集)
x, y = data.iloc[:, :-1], data['y']
# 打乱数据集。  su:  sklearn.utils
# random_state:随机种子。 若两次随机操作使用的随机种子相同,则随机结果也相同。
x, y = su.shuffle(x, y, random_state=7)
# 拆分测试集、训练集。
train_size = int(len(x) * 0.9)
train_x, test_x, train_y, test_y = \
    x[:train_size], x[train_size:], y[:train_size], y[train_size:]
train_x.shape, test_x.shape, train_y.shape, test_y.shape
((455, 13), (51, 13), (455,), (51,))

决策树回归

import sklearn.tree as st
import sklearn.metrics as sm
# 决策树回归模型,使用训练集训练模型,用测试集测试。
model = st.DecisionTreeRegressor(max_depth=4)
model.fit(train_x, train_y)  # 针对训练集数据进行训练
pred_test_y = model.predict(test_x)  # 针对测试集数据进行测试
print(sm.r2_score(test_y, pred_test_y))
print(sm.mean_absolute_error(test_y, pred_test_y))

# 输出单棵决策树的特征重要性指标
fi = model.feature_importances_
print(fi)
print(header)

0.678829897847864
2.923060066720631
[0.04992925 0.         0.         0.         0.02788393 0.62987544
 0.00067595 0.0908006  0.         0.         0.00415868 0.
 0.19667614]
['CRIM' 'ZN' 'INDUS' 'CHAS' 'NOX' 'RM' 'AGE' 'DIS' 'RAD' 'TAX' 'PTRATIO'
 'B' 'LSTAT']
import graphviz
dot_data = st.export_graphviz(model,out_file=None,feature_names= header)
graph = graphviz.Source(dot_data)
graph

在这里插入图片描述

正向激励

import sklearn.ensemble as se

# AdaBoost回归模型,使用训练集训练模型,用测试集测试。
# 调整参数
# params1 = [6, 7]
# params2 = [50, 100, 150, 200]

# for p1 in params1:
#     for p2 in params2:
#         model = st.DecisionTreeRegressor(max_depth=p1)
#         model = se.AdaBoostRegressor(model, n_estimators=p2, random_state=7)
#         model.fit(train_x, train_y)  # 针对训练集数据进行训练
#         pred_test_y = model.predict(test_x)  # 针对测试集数据进行测试
#         print('max_depth:', p1, '  n_estimators:', p2)
#         print(sm.r2_score(test_y, pred_test_y))
#         print(sm.mean_absolute_error(test_y, pred_test_y))

model = st.DecisionTreeRegressor(max_depth=6)
model = se.AdaBoostRegressor(model, n_estimators=100, random_state=7)
model.fit(train_x, train_y)  # 针对训练集数据进行训练
pred_test_y = model.predict(test_x)  # 针对测试集数据进行测试
print(sm.r2_score(test_y, pred_test_y))
print(sm.mean_absolute_error(test_y, pred_test_y))       
# 输出单棵决策树的特征重要性指标
fi = model.feature_importances_
print(fi)
print(header)
# 特征重要性可视化
s = pd.Series(fi, index=header)
s.sort_values().plot.barh()
0.8814458900728809
2.060418077972981
[0.03474378 0.00092137 0.00581656 0.00102895 0.03494231 0.23408824
 0.01790326 0.10480302 0.00889913 0.02894075 0.03176375 0.01503917
 0.48110971]
['CRIM' 'ZN' 'INDUS' 'CHAS' 'NOX' 'RM' 'AGE' 'DIS' 'RAD' 'TAX' 'PTRATIO'
 'B' 'LSTAT']


<matplotlib.axes._subplots.AxesSubplot at 0x2d786e57400>

在这里插入图片描述

GBDT

model = se.GradientBoostingRegressor(
    max_depth=6, n_estimators=1000, min_samples_split=15)
model.fit(train_x, train_y)
pred_test_y = model.predict(test_x)
print(sm.r2_score(test_y, pred_test_y))
print(sm.mean_absolute_error(test_y, pred_test_y))
0.9293816082262505
1.610908989775633

随机森林

model = se.RandomForestRegressor(
    max_depth=10, n_estimators=1000, min_samples_split=4)
model.fit(train_x, train_y)
pred_test_y = model.predict(test_x)
print(sm.r2_score(test_y, pred_test_y))
print(sm.mean_absolute_error(test_y, pred_test_y))
0.8977117124659044
1.9161244706102394

案例:共享单车投放量分析与预测

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import sklearn.preprocessing as sp
import sklearn.utils as su
import sklearn.linear_model as lm
import sklearn.tree as st
import sklearn.ensemble as se
import sklearn.metrics as sm
# 加载共享单车数据集,完成基本数据分析
data = pd.read_csv('../data/bike_day.csv')
# 特征工程,删除不必要的特征
data = data.drop(['instant', 'dteday', 'casual', 'registered'], axis=1)
data.describe()
seasonyrmnthholidayweekdayworkingdayweathersittempatemphumwindspeedcnt
count731.000000731.000000731.000000731.000000731.000000731.000000731.000000731.000000731.000000731.000000731.000000731.000000
mean2.4965800.5006846.5198360.0287282.9972640.6839951.3953490.4953850.4743540.6278940.1904864504.348837
std1.1108070.5003423.4519130.1671552.0047870.4652330.5448940.1830510.1629610.1424290.0774981937.211452
min1.0000000.0000001.0000000.0000000.0000000.0000001.0000000.0591300.0790700.0000000.02239222.000000
25%2.0000000.0000004.0000000.0000001.0000000.0000001.0000000.3370830.3378420.5200000.1349503152.000000
50%3.0000001.0000007.0000000.0000003.0000001.0000001.0000000.4983330.4867330.6266670.1809754548.000000
75%3.0000001.00000010.0000000.0000005.0000001.0000002.0000000.6554170.6086020.7302090.2332145956.000000
max4.0000001.00000012.0000001.0000006.0000001.0000003.0000000.8616670.8408960.9725000.5074638714.000000
# 离散型特征数值分析
data['weathersit'].value_counts()
1    463
2    247
3     21
Name: weathersit, dtype: int64
data.pivot_table(index='weathersit', values='cnt')
cnt
weathersit
14876.786177
24035.862348
31803.285714
data.pivot_table(index='workingday', values='cnt')
cnt
workingday
04330.168831
14584.820000
data.pivot_table(index='weekday', values='cnt')
cnt
weekday
04228.828571
14338.123810
24510.663462
34548.538462
44667.259615
54690.288462
64550.542857
# 连续型特征数值分析
data[['temp', 'atemp', 'hum', 'windspeed']].plot.box()
<matplotlib.axes._subplots.AxesSubplot at 0x2161e29dfd0>

在这里插入图片描述

训练机器学习模型,预测共享单车使用量
# 1. 整理输入集x 与 输出集y
x, y = data.iloc[:, :-1], data['cnt']
x.shape, y.shape

# 2. 拆分测试集与训练集
x, y = su.shuffle(x, y, random_state=7)
train_size = int(len(x) * 0.9)
train_x, test_x, train_y, test_y = \
    x[:train_size], x[train_size:], y[:train_size], y[train_size:]

# 3. 选择模型,训练模型
model = se.RandomForestRegressor(
    max_depth=10, n_estimators=1000, min_samples_split=3)
model.fit(train_x, train_y)

# 预测
pred_train_y = model.predict(train_x)
pred_test_y = model.predict(test_x)

# 4. 评估模型
print('train r2 score:', sm.r2_score(train_y, pred_train_y))
print('test  r2 score:', sm.r2_score(test_y, pred_test_y))
train r2 score: 0.9773682034502736
test  r2 score: 0.891479300161597
# 输出特征重要性
fi = model.feature_importances_
header = x.columns
s = pd.Series(fi, index=header)
s.sort_values().plot.barh()
<matplotlib.axes._subplots.AxesSubplot at 0x216204655c0>

在这里插入图片描述

以hour为单位的共享单车投放量预测
# 加载共享单车数据集,完成基本数据分析
data = pd.read_csv('../data/bike_hour.csv')
# 特征工程,删除不必要的特征
data = data.drop(['instant', 'dteday', 'casual', 'registered'], axis=1)
# 1. 整理输入集x 与 输出集y
x, y = data.iloc[:, :-1], data['cnt']
x.shape, y.shape

# 2. 拆分测试集与训练集
x, y = su.shuffle(x, y, random_state=7)
train_size = int(len(x) * 0.9)
train_x, test_x, train_y, test_y = \
    x[:train_size], x[train_size:], y[:train_size], y[train_size:]

# 3. 选择模型,训练模型
model = se.RandomForestRegressor(
    max_depth=10, n_estimators=1000, min_samples_split=3)
model.fit(train_x, train_y)

# 预测
pred_train_y = model.predict(train_x)
pred_test_y = model.predict(test_x)

# 4. 评估模型
print('train r2 score:', sm.r2_score(train_y, pred_train_y))
print('test  r2 score:', sm.r2_score(test_y, pred_test_y))
train r2 score: 0.9420587049496375
test  r2 score: 0.9185966492961931
# 输出特征重要性
fi = model.feature_importances_
header = x.columns
s = pd.Series(fi, index=header)
s.sort_values().plot.barh()
<matplotlib.axes._subplots.AxesSubplot at 0x21620685550>

在这里插入图片描述

分类模型:鸢尾花数据集

import numpy as np
import pandas as pd
import sklearn.datasets as sd
# 加载鸢尾花数据集
iris = sd.load_iris()
# 简单封装并分析
data = pd.DataFrame(iris.data, columns=iris.feature_names)
data['target'] = iris.target
data.describe()
sepal length (cm)sepal width (cm)petal length (cm)petal width (cm)target
count150.000000150.000000150.000000150.000000150.000000
mean5.8433333.0540003.7586671.1986671.000000
std0.8280660.4335941.7644200.7631610.819232
min4.3000002.0000001.0000000.1000000.000000
25%5.1000002.8000001.6000000.3000000.000000
50%5.8000003.0000004.3500001.3000001.000000
75%6.4000003.3000005.1000001.8000002.000000
max7.9000004.4000006.9000002.5000002.000000
# 分析不同类别下鸢尾花其他属性特征的规律
data.pivot_table(index='target')
petal length (cm)petal width (cm)sepal length (cm)sepal width (cm)
target
01.4640.2445.0063.418
14.2601.3265.9362.770
25.5522.0266.5882.974
data.plot.scatter(x='sepal length (cm)', y='sepal width (cm)', 
                 c='target', cmap='brg')
<matplotlib.axes._subplots.AxesSubplot at 0x17de30ff080>

在这里插入图片描述

data.plot.scatter(x='petal length (cm)', y='petal width (cm)', 
                 c='target', cmap='brg')
<matplotlib.axes._subplots.AxesSubplot at 0x17de324ff28>

在这里插入图片描述

# 整理一组简单数据,训练逻辑回归模型,完成分类业务。
import sklearn.utils as su
import sklearn.linear_model as lm

sub_data = data.tail(150)
# 整理输入集与输出集
x, y = sub_data.iloc[:, :-1], sub_data['target']
# 拆分测试集与训练集
x, y = su.shuffle(x, y, random_state=7)
train_size = 70
train_x, test_x, train_y, test_y = \
    x[:train_size], x[train_size:], y[:train_size], y[train_size:]
# 训练模型,并使用测试数据评估模型
model = lm.LogisticRegression()
model.fit(train_x, train_y)
pred_test_y = model.predict(test_x)
# 对于分类结果的评估:
(pred_test_y==test_y).sum() / test_y.size
0.975
import matplotlib.pyplot as plt
x = np.linspace(-10, 10, 200)
y = 1 / (1 + np.exp(-x))
plt.grid(linestyle=':')
plt.plot(x, y)
[<matplotlib.lines.Line2D at 0x17de33d4358>]

在这里插入图片描述

data[['petal length (cm)', 'target']].head(100).plot.scatter(x='petal length (cm)', y='target')
<matplotlib.axes._subplots.AxesSubplot at 0x17de33fd128>

在这里插入图片描述

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值