多项式回归模型做运费价格预测demo

多项式回归模型做运费价格预测

数据准备

数据源取系统应付结算单数据，数据需要经过清洗，具体为：

SELECT
	b.CODE,
	b.TRAVEL_MILES,
	b.WEIGHT,
	b.VOLUME,
	b.PS_NUM 
FROM
	base_settle b 
WHERE
	b.TRAVEL_MILES > 0 
	AND WEIGHT > 0 
	AND VOLUME > 0 
	AND PS_NUM > 0
	and code like 'AP%';

导包并读取数据源

import csv
import numpy as np
import pandas as pd
import os
source=pd.read_csv('C:/Users/liuhb41/Desktop/回归/settledata1.csv')
source

分割训练参数及结果

#分割X和Y
from sklearn.model_selection import train_test_split 
data_train, data_test = train_test_split(source, test_size=0.2, random_state=1234)

设置训练集X和Y

x_train=data_train.iloc[:,2:5]
x_train

将x转换为ndarray

x_train=x_train.values
x_train

设置训练集y

y_train=data_train['HJ_FEE']
y_train

进行数据多项式处理

from sklearn.preprocessing import PolynomialFeatures
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
d=3#多项式维度
rnd=np.random.RandomState(40)#设置随机数种子
x_train_p=PolynomialFeatures(degree=d,interaction_only=False).fit_transform(x_train)#对数据进行多项式处理

模型创建及预测

#创建线性回归模型
LinearR_=LinearRegression()
#模型预测
LinearR_.fit(x_train_p, y_train)

模型测试

#区分测试集X,Y
data_test
x_test=data_test.iloc[:,2:5]
y_test=data_test['HJ_FEE']
#预测
y_test_predict=LinearR_.predict(PolynomialFeatures(degree=d).fit_transform(x_test))

可用score函数查看评分，不断调优

score=LinearR_.score(x_train_p, y_train)
score