Datawhale 数据挖掘入门：建模调参 笔记（3）

TASK4:建模调参
摘自 小雨姑娘 在Datawhale 数据挖掘入门：建模调参部分的讲义

1 线性回归

假设已经有了数据train_X，和trian_y
模型的建立

from sklearn.linear_model import LinearRegression
model = LinearRegression(normalize=True)
model = model.fit(train_X, train_y)

查看训练的线性回归模型的截距（intercept）与权重(coef)

'intercept:'+ str(model.intercept_)
sorted(dict(zip(continuous_feature_names, model.coef_)).items(), key=lambda x:x[1], reverse=True)

绘制某一个特征与标签的散点图，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

from matplotlib import pyplot as plt
subsample_index = np.random.randint(low=0, high=len(train_y), size=50)
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()

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。

import seaborn as sns
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)])

对标签进行了 $log(x+1)$ 变换，使标签贴近于正态分布。

train_y_ln = np.log(train_y + 1)

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)])

再次构建模型：

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)

可视化确认数据是否正常

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()

注：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）。

2.五折交叉验证

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)))
        return result
    return wrapper

对原始标签train_y进行五折交叉验证过程

scores = cross_val_score(model, X=train_X, y=train_y, verbose=1, cv = 5, scoring=make_scorer(log_transfer(mean_absolute_error)))

计算平均绝对误差的平均值

print('AVG:', np.mean(scores))

对修正过的标签train_y_ln:

scores = cross_val_score(model, X=train_X, y=train_y_ln, verbose=1, cv = 5, scoring=make_scorer(mean_absolute_error))

print('AVG:', np.mean(scores))

显示每一折的MAE

scores = pd.DataFrame(scores.reshape(1,-1))
scores.columns = ['cv' + str(x) for x in range(1, 6)]
scores.index = ['MAE']
scores

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

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)

model = model.fit(train_X, train_y_ln)
mean_absolute_error(val_y_ln, model.predict(val_X))

学习曲线和学习率的绘制

from sklearn.model_selection import learning_curve, validation_curve

def plot_learning_curve(estimator, title, X, y, ylim=None, cv=None,n_jobs=1, train_size=np.linspace(.1, 1.0, 5 )):  
    plt.figure()  
    plt.title(title)  
    if ylim is not None:  
        plt.ylim(*ylim)  
    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")  
    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")  
    plt.plot(train_sizes, test_scores_mean,'o-',color="g",  
             label="Cross-validation score")  
    plt.legend(loc="best")  
    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)  

3. 不同模型的对比

载入数据

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)

续前特征选择：
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()]
          

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')

对比三种线性回归的效果：

result = pd.DataFrame(result)
result.index = ['cv' + str(x) for x in range(1, 6)]
result

线性回归中特征向量对最终结果的影响：

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正则化对特征向量的影响
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。

###4 非线性模型

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 LGBMRegressor

models = [LinearRegression(),
          DecisionTreeRegressor(),
          RandomForestRegressor(),
          GradientBoostingRegressor(),
          MLPRegressor(solver='lbfgs', max_iter=100), 
          XGBRegressor(n_estimators = 100, objective='reg:squarederror'), 
          LGBMRegressor(n_estimators = 100)]

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')

罗列每一个模型每一折的MAE

result = pd.DataFrame(result)
result.index = ['cv' + str(x) for x in range(1, 6)]
result

4 参数的调节

4.1 贪心调参

## LGB的参数集合：

objective = ['regression', 'regression_l1', 'mape', 'huber', 'fair']

num_leaves = [3,5,10,15,20,40, 55]
max_depth = [3,5,10,15,20,40, 55]
bagging_fraction = []
feature_fraction = []
drop_rate = []

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

绘制曲线：

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())])

4.2 网格搜索

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)))

4.3 贝叶斯调参方法

from bayes_opt import BayesianOptimization
def 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 - val

rf_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']

5 总结

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]

绘制所有的改进所带来的的精度的提升。可以发现，对数变换带来的变化最为显著，再一次说明了特征选择上下功夫的重要性，参数调节所带来的的提升是相对有限的。