基于决策树（Decision Tree）的bagging算法：随机森林（Random Forest）（包括具体代码）

最新推荐文章于 2022-04-22 21:10:44 发布

原创

最新推荐文章于 2022-04-22 21:10:44 发布 · 3.1k 阅读

12 ·

CC 4.0 BY-SA版权

重点先跑模型：
数据：
http://sofasofa.io/competition.php?id=1
代码：
需要修改一下文件的地址：
代码运行的时候一步一步来：

# -*- coding: utf-8 -*-
# import 常用包
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# 数据完整化和数据数值化
from sklearn.preprocessing import Imputer
from sklearn.preprocessing import LabelEncoder

# 特征提取和特征选择
from sklearn.feature_selection import f_regression
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
from sklearn.decomposition import PCA

# 数据归一化和数据标准化
from sklearn.preprocessing import StandardScaler
from sklearn.preprocessing import Normalizer

# 数据分割
from sklearn.model_selection import train_test_split
from sklearn.model_selection import KFold

# 模型构建
from sklearn.ensemble import RandomForestRegressor
import lightgbm as lgb

# 误差评估
from sklearn.metrics import mean_squared_error
from sklearn.metrics import mean_absolute_error

# 参数调节
from sklearn.model_selection import GridSearchCV

# -----------------------------------------------加载数据，数据处理，特征提取，特征选择------------------------------------------------------
# 加载数据
train = pd.read_csv('E:/dataset/reg/bike/train.csv')
test = pd.read_csv('E:/dataset/reg/bike/test.csv')
submit = pd.read_csv('E:/dataset/reg/bike/sample_submit.csv')

# ----------------------------------------------------------------------数据分割和模型使用-------------------------------------------------------------------------
predictors = ['city', 'hour', 'is_workday', 'weather', 'temp_1', 'temp_2', 'wind']
X_train, X_test, Y_train, Y_test = train_test_split(train[predictors], train["y"], test_size=0.2)
model = RandomForestRegres