學(xué)習(xí)目標(biāo)
- 了解常用的機(jī)器學(xué)習(xí)模型霞怀，并掌握機(jī)器學(xué)習(xí)模型的建模與調(diào)參流程
內(nèi)容介紹
- 線性回歸模型：
- 線性回歸對于特征的要求吉殃；
- 處理長尾分布酱鸭；
- 理解線性回歸模型劫樟；
- 模型性能驗證：
- 評價函數(shù)與目標(biāo)函數(shù)痪枫；
- 交叉驗證方法；
- 留一驗證方法叠艳；
- 針對時間序列問題的驗證奶陈；
- 繪制學(xué)習(xí)率曲線；
- 繪制驗證曲線附较；
- 嵌入式特征選擇：
- Lasso回歸吃粒；
- Ridge回歸；
- 決策樹翅睛；
- 模型對比：
- 常用線性模型声搁；
- 常用非線性模型黑竞；
- 模型調(diào)參：
- 貪心調(diào)參方法捕发；
- 網(wǎng)格調(diào)參方法；
- 貝葉斯調(diào)參方法很魂；
本文推薦了一些博客與教材供初學(xué)者們進(jìn)行學(xué)習(xí)扎酷。
線性回歸模型
https://zhuanlan.zhihu.com/p/49480391
決策樹模型
https://zhuanlan.zhihu.com/p/65304798
GBDT模型
https://zhuanlan.zhihu.com/p/45145899
XGBoost模型
https://zhuanlan.zhihu.com/p/86816771
LightGBM模型
https://zhuanlan.zhihu.com/p/89360721
推薦教材：
- 《機(jī)器學(xué)習(xí)》 https://book.douban.com/subject/26708119/
- 《統(tǒng)計學(xué)習(xí)方法》 https://book.douban.com/subject/10590856/
- 《Python大戰(zhàn)機(jī)器學(xué)習(xí)》 https://book.douban.com/subject/26987890/
- 《面向機(jī)器學(xué)習(xí)的特征工程》 https://book.douban.com/subject/26826639/
- 《數(shù)據(jù)科學(xué)家訪談錄》 https://book.douban.com/subject/30129410/
代碼示例
讀取數(shù)據(jù)
import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore')
reduce_mem_usage 函數(shù)通過調(diào)整數(shù)據(jù)類型，幫助我們減少數(shù)據(jù)在內(nèi)存中占用的空間
def reduce_mem_usage(df):
""" iterate through all the columns of a dataframe and modify the data type
to reduce memory usage.
"""
start_mem = df.memory_usage().sum()
print('Memory usage of dataframe is {:.2f} MB'.format(start_mem))
for col in df.columns:
col_type = df[col].dtype
if col_type != object:
c_min = df[col].min()
c_max = df[col].max()
if str(col_type)[:3] == 'int':
if c_min > np.iinfo(np.int8).min and c_max < np.iinfo(np.int8).max:
df[col] = df[col].astype(np.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:
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:
df[col] = df[col].astype('category')
end_mem = df.memory_usage().sum()
print('Memory usage after optimization is: {:.2f} MB'.format(end_mem))
print('Decreased by {:.1f}%'.format(100 * (start_mem - end_mem) / start_mem))
return df
sample_feature = reduce_mem_usage(pd.read_csv('data_for_tree.csv'))
Memory usage of dataframe is 60507328.00 MB
Memory usage after optimization is: 15724107.00 MB
Decreased by 74.0%
continuous_feature_names = [x for x in sample_feature.columns if x not in ['price','brand','model','brand']]
線性回歸 & 五折交叉驗證 & 模擬真實業(yè)務(wù)情況
sample_feature = sample_feature.dropna().replace('-', 0).reset_index(drop=True)
sample_feature['notRepairedDamage'] = sample_feature['notRepairedDamage'].astype(np.float32)
train = sample_feature[continuous_feature_names + ['price']]
train_X = train[continuous_feature_names]
train_y = train['price']
建模
from sklearn.linear_model import LinearRegression
model = LinearRegression(normalize=True)
model = model.fit(train_X, train_y)
查看訓(xùn)練的線性回歸模型的截距（intercept）與權(quán)重(coef)
'intercept:'+ str(model.intercept_)
sorted(dict(zip(continuous_feature_names, model.coef_)).items(), key=lambda x:x[1], reverse=True)
[('v_6', 3342612.384537345),
('v_8', 684205.534533214),
('v_9', 178967.94192530424),
('v_7', 35223.07319016895),
('v_5', 21917.550249749802),
('v_3', 12782.03250792227),
('v_12', 11654.925634146672),
('v_13', 9884.194615297649),
('v_11', 5519.182176035517),
('v_10', 3765.6101415594258),
('gearbox', 900.3205339198406),
('fuelType', 353.5206495542567),
('bodyType', 186.51797317460046),
('city', 45.17354204168846),
('power', 31.163045441455335),
('brand_price_median', 0.535967111869784),
('brand_price_std', 0.4346788365040235),
('brand_amount', 0.15308295553300566),
('brand_price_max', 0.003891831020467389),
('seller', -1.2684613466262817e-06),
('offerType', -4.759058356285095e-06),
('brand_price_sum', -2.2430642281682917e-05),
('name', -0.00042591632723759166),
('used_time', -0.012574429533889028),
('brand_price_average', -0.414105722833381),
('brand_price_min', -2.3163823428971835),
('train', -5.392535065078232),
('power_bin', -59.24591853031839),
('v_14', -233.1604256172217),
('kilometer', -372.96600915402496),
('notRepairedDamage', -449.29703564695365),
('v_0', -1490.6790578168238),
('v_4', -14219.648899108111),
('v_2', -16528.55239086934),
('v_1', -42869.43976200439)]
from matplotlib import pyplot as plt
subsample_index = np.random.randint(low=0, high=len(train_y), size=50)
繪制特征v_9的值與標(biāo)簽的散點圖遏匆，圖片發(fā)現(xiàn)模型的預(yù)測結(jié)果（藍(lán)色點）與真實標(biāo)簽（黑色點）的分布差異較大法挨，且部分預(yù)測值出現(xiàn)了小于0的情況，說明我們的模型存在一些問題
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()
The predicted price is obvious different from true price
通過作圖我們發(fā)現(xiàn)數(shù)據(jù)的標(biāo)簽（price）呈現(xiàn)長尾分布幅聘，不利于我們的建模預(yù)測凡纳。原因是很多模型都假設(shè)數(shù)據(jù)誤差項符合正態(tài)分布，而長尾分布的數(shù)據(jù)違背了這一假設(shè)帝蒿。參考博客：https://blog.csdn.net/Noob_daniel/article/details/76087829
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)])
It is clear to see the price shows a typical exponential distribution
<matplotlib.axes._subplots.AxesSubplot at 0x1b33efb2f98>
在這里我們對標(biāo)簽進(jìn)行了 變換荐糜，使標(biāo)簽貼近于正態(tài)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)])
The transformed price seems like normal distribution
<matplotlib.axes._subplots.AxesSubplot at 0x1b33f077160>
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)
intercept:23.515920686637713
[('v_9', 6.043993029165403),
('v_12', 2.0357439855551394),
('v_11', 1.3607608712255672),
('v_1', 1.3079816298861897),
('v_13', 1.0788833838535354),
('v_3', 0.9895814429387444),
('gearbox', 0.009170812023421397),
('fuelType', 0.006447089787635784),
('bodyType', 0.004815242907679581),
('power_bin', 0.003151801949447194),
('power', 0.0012550361843629999),
('train', 0.0001429273782925814),
('brand_price_min', 2.0721302299502698e-05),
('brand_price_average', 5.308179717783439e-06),
('brand_amount', 2.8308531339942507e-06),
('brand_price_max', 6.764442596115763e-07),
('offerType', 1.6765966392995324e-10),
('seller', 9.308109838457312e-12),
('brand_price_sum', -1.3473184925468486e-10),
('name', -7.11403461065247e-08),
('brand_price_median', -1.7608143661053008e-06),
('brand_price_std', -2.7899058266986454e-06),
('used_time', -5.6142735899344175e-06),
('city', -0.0024992974087053223),
('v_14', -0.012754139659375262),
('kilometer', -0.013999175312751872),
('v_0', -0.04553774829634237),
('notRepairedDamage', -0.273686961116076),
('v_7', -0.7455902679730504),
('v_4', -0.9281349233755761),
('v_2', -1.2781892166433606),
('v_5', -1.5458846136756323),
('v_10', -1.8059217242413748),
('v_8', -42.611729973490604),
('v_6', -241.30992120503035)]
再次進(jìn)行可視化，發(fā)現(xiàn)預(yù)測結(jié)果與真實值較為接近葛超，且未出現(xiàn)異常狀況
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()
The predicted price seems normal after np.log transforming
模擬真實業(yè)務(wù)情況
但在事實上暴氏，由于我們并不具有預(yù)知未來的能力，五折交叉驗證在某些與時間相關(guān)的數(shù)據(jù)集上反而反映了不真實的情況绣张。通過2018年的二手車價格預(yù)測2017年的二手車價格答渔，這顯然是不合理的，因此我們還可以采用時間順序?qū)?shù)據(jù)集進(jìn)行分隔侥涵。在本例中沼撕，我們選用靠前時間的4/5樣本當(dāng)作訓(xùn)練集宋雏，靠后時間的1/5當(dāng)作驗證集，最終結(jié)果與五折交叉驗證差距不大
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))
0.19443858353490887
除此之外务豺，決策樹通過信息熵或GINI指數(shù)選擇分裂節(jié)點時好芭，優(yōu)先選擇的分裂特征也更加重要，這同樣是一種特征選擇的方法冲呢。XGBoost與LightGBM模型中的model_importance指標(biāo)正是基于此計算的
模型調(diào)參
在此我們介紹了三種常用的調(diào)參方法如下：
- 貪心算法 http://www.reibang.com/p/ab89df9759c8
- 網(wǎng)格調(diào)參 https://blog.csdn.net/weixin_43172660/article/details/83032029
- 貝葉斯調(diào)參 https://blog.csdn.net/linxid/article/details/81189154
## LGB的參數(shù)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 = []
貪心調(diào)參
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())])
<matplotlib.axes._subplots.AxesSubplot at 0x1fea93f6080>
Grid Search 調(diào)參
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_
{'max_depth': 15, 'num_leaves': 55, 'objective': 'regression'}
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)))
0.13626164479243302
總結(jié)
在本章中舍败，我們完成了建模與調(diào)參的工作，并對我們的模型進(jìn)行了驗證敬拓。此外邻薯，我們還采用了一些基本方法來提高預(yù)測的精度，提升如下圖所示乘凸。
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])
Task5 建模調(diào)參 END.
