读取数据
导入相关包并读取数据
import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore')类型，节省内存
def reduce_mem_usage(df):
""" iterate through all the columns of a dataframe and modify the data type
to reduce memory usage. 通过调整数据类型，帮助我们减少数据在内存中占用的空间
"""
start_mem = df.memory_usage().sum()#获取数据内存使用情况
print('Memory usage of dataframe is {:.2f} MB'.format(start_mem))
for col in df.columns:
col_type = df[col].dtype #获取某一列的数据类型
if col_type != object: #若该列不是对象类型
c_min = df[col].min()
c_max = df[col].max() #获取到该列的最大最小值
if str(col_type)[:3] == 'int': #若该数据类型是intxxx
if c_min > np.iinfo(np.int8).min and c_max < np.iinfo(np.int8).max:
df[col] = df[col].astype(np.int8) #若数据是在int8范围内，则改变数据类型为int8
elif c_min > np.iinfo(np.int16).min and c_max < np.iinfo(np.int16).max:
df[col] = df[col].astype(np.int16)
elif c_min > np.iinfo(np.int32).min and c_max < np.iinfo(np.int32).max:
df[col] = df[col].astype(np.int32)
elif c_min > np.iinfo(np.int64).min and c_max < np.iinfo(np.int64).max:
df[col] = df[col].astype(np.int64)
else: #不是intxxx，则为floatxxx
if c_min > np.finfo(np.float16).min and c_max < np.finfo(np.float16).max:
df[col] = df[col].astype(np.float16)
elif c_min > np.finfo(np.float32).min and c_max < np.finfo(np.float32).max:
df[col] = df[col].astype(np.float32)
else:
df[col] = df[col].astype(np.float64)
else: #若数据类型是object类型的，则设置为category
df[col] = df[col].astype('category')
end_mem = df.memory_usage().sum()
print('Memory usage after optimization is: {:.2f} MB'.format(end_mem)) #现在的数据内存使用情况
print('Decreased by {:.1f}%'.format(100 * (start_mem - end_mem) / start_mem)) #降低内存使用百分比
return df读取之前特征工程生成的数据
sample_feature = reduce_mem_usage(pd.read_csv('data_for_tree.csv')) #data_for_tree是之前特征工程生成的
continuous_feature_names = [x for x in sample_feature.columns if x not in ['price','brand','model']]#获取除价格、品牌、车型编码的属性名
线性回归 & 五折交叉验证 & 模拟真实业务情况
简单处理NAN，生成特征和标签
sample_feature = sample_feature.dropna().replace('-', 0).reset_index(drop=True) #删除nan,将-替代为0,并重置索引,直接简单删除了NAN
sample_feature['notRepairedDamage'] = sample_feature['notRepairedDamage'].astype(np.float32) #原本是category的
train = sample_feature[continuous_feature_names + ['price']] #训练集是以上的加上price
train_X = train[continuous_feature_names]
train_y = train['price']简单建模
from sklearn.linear_model import LinearRegression
model = LinearRegression(normalize=True)
model = model.fit(train_X, train_y)print('intercept:'+ str(model.intercept_))
sorted(dict(zip(continuous_feature_names, model.coef_)).items(), key=lambda x:x[1], reverse=True) #查看训练的线性回归模型的截距（intercept）与权重(coef)
from matplotlib import pyplot as plt
subsample_index = np.random.randint(low=0, high=len(train_y), size=50) #随机采样
#绘制特征v_9的值与标签的散点图，图片发现模型的预测结果（蓝色点）与真实标签（黑色点）的分布差异较大，
#且部分预测值出现了小于0的情况，说明我们的模型存在一些问题
plt.scatter(train_X['v_9'][subsample_index], train_y[subsample_index], color='black')
plt.scatter(train_X['v_9'][subsample_index], model.predict(train_X.loc[subsample_index]), color='blue')
plt.xlabel('v_9')
plt.ylabel('price')
plt.legend(['True Price','Predicted Price'],loc='upper right')
print('The predicted price is obvious different from true price')
plt.show()
import seaborn as sns
#通过作图我们发现数据的标签（price）呈现长尾分布，不利于我们的建模预测。
#原因是很多模型都假设数据误差项符合正态分布，而长尾分布的数据违背了这一假设
print('It is clear to see the price shows a typical exponential distribution')
plt.figure(figsize=(15,5))
plt.subplot(1,2,1)
sns.distplot(train_y)
plt.subplot(1,2,2)
sns.distplot(train_y[train_y < np.quantile(train_y, 0.9)]) #np.quantile(train_y, 0.9)train_y的0.9分位数，
使标签符合正态分布
train_y_ln = np.log(train_y + 1) #在这里我们对标签进行了 log(x+1)变换，使标签贴近于正态分布
import seaborn as sns
print('The transformed price seems like normal distribution')
plt.figure(figsize=(15,5))
plt.subplot(1,2,1)
sns.distplot(train_y_ln)
plt.subplot(1,2,2)
sns.distplot(train_y_ln[train_y_ln < np.quantile(train_y_ln, 0.9)])
model = model.fit(train_X, train_y_ln)
print('intercept:'+ str(model.intercept_))
sorted(dict(zip(continuous_feature_names, model.coef_)).items(), key=lambda x:x[1], reverse=True)
#再次进行可视化，发现预测结果与真实值较为接近，且未出现异常状况
plt.scatter(train_X['v_9'][subsample_index], train_y[subsample_index], color='black')
plt.scatter(train_X['v_9'][subsample_index], np.exp(model.predict(train_X.loc[subsample_index])), color='blue')
plt.xlabel('v_9')
plt.ylabel('price')
plt.legend(['True Price','Predicted Price'],loc='upper right')
print('The predicted price seems normal after np.log transforming')
plt.show()
五折交叉验证
在使用训练集对参数进行训练的时候，经常会发现人们通常会将一整个训练集分为三个部分。一般分为：训练集（train_set），评估集（valid_set），
测试集（test_set）这三个部分。这其实是为了保证训练效果而特意设置的。其中测试集很好理解，其实就是完全不参与训练的数据，仅仅用来观测测试效果
的数据。而训练集和评估集则牵涉到下面的知识了。
因为在实际的训练中，训练的结果对于训练集的拟合程度通常还是挺好的（初始条件敏感），但是对于训练集之外的数据的拟合程度通常就不那么令人满意了。
因此我们通常并不会把所有的数据集都拿来训练，而是分出一部分来（这一部分不参加训练）对训练集生成的参数进行测试，相对客观的判断这些参数对训练集
之外的数据的符合程度。这种思想就称为交叉验证（Cross Validation）
from sklearn.model_selection import cross_val_score
from sklearn.metrics import mean_absolute_error, make_scorer
def log_transfer(func):
def wrapper(y, yhat):
result = func(np.log(y), np.nan_to_num(np.log(yhat))) #np.nan_to_num()使用0代替数组x中的nan元素，使用有限的数字代替inf元素
return result
return wrapperscores = cross_val_score(model, X=train_X, y=train_y, verbose=1, cv = 5, scoring=make_scorer(log_transfer(mean_absolute_error)))#使用线性回归模型，对未处理标签的特征数据进行五折交叉验证
print('AVG:', np.mean(scores)) 
scores = cross_val_score(model, X=train_X, y=train_y_ln, verbose=1, cv = 5, scoring=make_scorer(mean_absolute_error)) #使用线性回归模型，对处理过标签的特征数据进行五折交叉验证
print('AVG:', np.mean(scores))
scores = pd.DataFrame(scores.reshape(1,-1))
scores.columns = ['cv' + str(x) for x in range(1, 6)]
scores.index = ['MAE']
scores
模拟真实业务情况
但在事实上，由于我们并不具有预知未来的能力，五折交叉验证在某些与时间相关的数据集上反而反映了不真实的情况。
通过2018年的二手车价格预测2017年的二手车价格，这显然是不合理的，因此我们还可以采用时间顺序对数据集进行分隔。
在本例中，我们选用靠前时间的4/5样本当作训练集，靠后时间的1/5当作验证集，最终结果与五折交叉验证差距不大
import datetime
sample_feature = sample_feature.reset_index(drop=True)
split_point = len(sample_feature) // 5 * 4 #取4/5的位置作为分隔点
train = sample_feature.loc[:split_point].dropna()
val = sample_feature.loc[split_point:].dropna()
train_X = train[continuous_feature_names]
train_y_ln = np.log(train['price'] + 1)
val_X = val[continuous_feature_names]
val_y_ln = np.log(val['price'] + 1) #生成训练集和验证集model = model.fit(train_X, train_y_ln)
mean_absolute_error(val_y_ln, model.predict(val_X))
绘制学习率曲线与验证曲线
from sklearn.model_selection import learning_curve, validation_curve
? learning_curve #？可以放在前面或者后面，查看这个东西的详情def plot_learning_curve(estimator, title, X, y, ylim=None, cv=None,n_jobs=1, train_size=np.linspace(.1, 1.0, 5 )): #绘制学习曲线
plt.figure()
plt.title(title)
if ylim is not None:
plt.ylim(*ylim) #设置y轴的数值显示范围
plt.xlabel('Training example')
plt.ylabel('score')
train_sizes, train_scores, test_scores = learning_curve(estimator, X, y, cv=cv, n_jobs=n_jobs, train_sizes=train_size, scoring = make_scorer(mean_absolute_error))
train_scores_mean = np.mean(train_scores, axis=1)
train_scores_std = np.std(train_scores, axis=1)
test_scores_mean = np.mean(test_scores, axis=1)
test_scores_std = np.std(test_scores, axis=1)
plt.grid()#区域
plt.fill_between(train_sizes, train_scores_mean - train_scores_std,
train_scores_mean + train_scores_std, alpha=0.1,
color="r") #绘制曲线的上下两部分，均值+标准差，均值-标准差
plt.fill_between(train_sizes, test_scores_mean - test_scores_std,
test_scores_mean + test_scores_std, alpha=0.1,
color="g")
plt.plot(train_sizes, train_scores_mean, 'o-', color='r',
label="Training score") #训练集尺寸和训练得分的平均分
plt.plot(train_sizes, test_scores_mean,'o-',color="g",
label="Cross-validation score")
plt.legend(loc="best") #给图加上图例
return plt plot_learning_curve(LinearRegression(), 'Liner_model', train_X[:1000], train_y_ln[:1000], ylim=(0.0, 0.5), cv=5, n_jobs=1) 
多种模型对比
train = sample_feature[continuous_feature_names + ['price']].dropna()
train_X = train[continuous_feature_names]
train_y = train['price']
train_y_ln = np.log(train_y + 1)线性模型 & 嵌入式特征选择
在过滤式和包裹式特征选择方法中，特征选择过程与学习器训练过程有明显的分别。而嵌入式特征选择在学习器训练过程中自动地进行特征选择。
嵌入式选择最常用的是L1正则化与L2正则化。在对线性回归模型加入两种正则化方法后，他们分别变成了岭回归与Lasso回归
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Ridge
from sklearn.linear_model import Lasso
models = [LinearRegression(),
Ridge(),
Lasso()] #共三种模型result = dict()
for model in models:
model_name = str(model).split('(')[0] #每一个按（分割，取第一个，即模型名称
scores = cross_val_score(model, X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error))
result[model_name] = scores
print(model_name + ' is finished')
result = pd.DataFrame(result) #三种方法的效果对比
result.index = ['cv' + str(x) for x in range(1, 6)]
result 
model = LinearRegression().fit(train_X, train_y_ln)
print('intercept:'+ str(model.intercept_))
sns.barplot(abs(model.coef_), continuous_feature_names)
L2正则化在拟合过程中通常都倾向于让权值尽可能小，最后构造一个所有参数都比较小的模型。因为一般认为参数值小的模型比较简单，
能适应不同的数据集，也在一定程度上避免了过拟合现象。可以设想一下对于一个线性回归方程，若参数很大，那么只要数据偏移一点点，
就会对结果造成很大的影响；但如果参数足够小，数据偏移得多一点也不会对结果造成什么影响，专业一点的说法是『抗扰动能力强』
model = Ridge().fit(train_X, train_y_ln)
print('intercept:'+ str(model.intercept_))
sns.barplot(abs(model.coef_), continuous_feature_names)
#L1正则化有助于生成一个稀疏权值矩阵，进而可以用于特征选择。如下图，我们发现power与userd_time特征非常重要
model = Lasso().fit(train_X, train_y_ln)
print('intercept:'+ str(model.intercept_))
sns.barplot(abs(model.coef_), continuous_feature_names)
除此之外，决策树通过信息熵或GINI指数选择分裂节点时，优先选择的分裂特征也更加重要，这同样是一种特征选择的方法。XGBoost与LightGBM模型中的model_importance指标正是基于此计算的
非线性模型
除了线性模型以外，还有许多我们常用的非线性模型如下，在此篇幅有限不再一一讲解原理。我们选择了部分常用模型与线性模型进行效果比对。
from sklearn.linear_model import LinearRegression
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.neural_network import MLPRegressor
from xgboost.sklearn import XGBRegressor
from lightgbm.sklearn import LGBMRegressormodels = [LinearRegression(),
DecisionTreeRegressor(),
RandomForestRegressor(),
GradientBoostingRegressor(),
MLPRegressor(solver='lbfgs', max_iter=100),
XGBRegressor(n_estimators = 100, objective='reg:squarederror'),
LGBMRegressor(n_estimators = 100)] #使用了线性回归，决策树回归，随机森林回归，GB回归，MLP回归，XGB回归，LGB回归result = dict()
for model in models:
model_name = str(model).split('(')[0]
scores = cross_val_score(model, X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error))
result[model_name] = scores
print(model_name + ' is finished')
result = pd.DataFrame(result)
result.index = ['cv' + str(x) for x in range(1, 6)]
result
模型调参
## LGB的参数集合：
objective = ['regression', 'regression_l1', 'mape', 'huber', 'fair']
num_leaves = [3,5,10,15,20,40, 55] #叶子数量
max_depth = [3,5,10,15,20,40, 55] #最大深度
bagging_fraction = []
feature_fraction = []
drop_rate = []贪心调参
best_obj = dict()
for obj in objective:
model = LGBMRegressor(objective=obj)
score = np.mean(cross_val_score(model, X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error)))
best_obj[obj] = score
best_leaves = dict()
for leaves in num_leaves:
model = LGBMRegressor(objective=min(best_obj.items(), key=lambda x:x[1])[0], num_leaves=leaves)
score = np.mean(cross_val_score(model, X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error)))
best_leaves[leaves] = score、
best_depth = dict()
for depth in max_depth:
model = LGBMRegressor(objective=min(best_obj.items(), key=lambda x:x[1])[0],
num_leaves=min(best_leaves.items(), key=lambda x:x[1])[0],
max_depth=depth)
score = np.mean(cross_val_score(model, X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error)))
best_depth[depth] = score sns.lineplot(x=['0_initial','1_turning_obj','2_turning_leaves','3_turning_depth'], y=[0.143 ,min(best_obj.values()), min(best_leaves.values()), min(best_depth.values())])
Grid Search 调参
from sklearn.model_selection import GridSearchCV
parameters = {'objective': objective , 'num_leaves': num_leaves, 'max_depth': max_depth}
model = LGBMRegressor()
clf = GridSearchCV(model, parameters, cv=5)
clf = clf.fit(train_X, train_y)
clf.best_params_
model = LGBMRegressor(objective='regression',
num_leaves=55,
max_depth=15)np.mean(cross_val_score(model, X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error)))
贝叶斯调参
贝叶斯优化是一种用模型找到函数最小值方法，已经应用于机器学习问题中的超参数搜索，这种方法性能好，同时比随机搜索省时。
贝叶斯方法与随机或网格搜索的不同之处在于，它在尝试下一组超参数时，会参考之前的评估结果，因此可以省去很多无用功
from bayes_opt import BayesianOptimizationdef rf_cv(num_leaves, max_depth, subsample, min_child_samples):
val = cross_val_score(
LGBMRegressor(objective = 'regression_l1',
num_leaves=int(num_leaves),
max_depth=int(max_depth),
subsample = subsample,
min_child_samples = int(min_child_samples)
),
X=train_X, y=train_y_ln, verbose=0, cv = 5, scoring=make_scorer(mean_absolute_error)
).mean()
return 1 - valrf_bo = BayesianOptimization(
rf_cv,
{
'num_leaves': (2, 100),
'max_depth': (2, 100),
'subsample': (0.1, 1),
'min_child_samples' : (2, 100)
}
)rf_bo.maximize()
1 - rf_bo.max['target']
总结
#在本章中，我们完成了建模与调参的工作，并对我们的模型进行了验证。此外，我们还采用了一些基本方法来提高预测的精度，提升如下图所示
plt.figure(figsize=(13,5))
sns.lineplot(x=['0_origin','1_log_transfer','2_L1_&_L2','3_change_model','4_parameter_turning'], y=[1.36 ,0.19, 0.19, 0.14, 0.13])
193
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
