预测销售量

前提
#认为天猫销量和年份之间存在一定函数关系
导包

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline


对年份进行处理

years &＃61; np.arange(2009,2020)
years


sales &＃61; np.array([0.5,9.36,52,191,352,571,912,1207,1682.69,2135,2684])
sales


plt.scatter(years,sales,c &＃61; &＃39;red&＃39;,marker &＃61; &＃39;*&＃39;,s &＃61; 20)


X &＃61; (years - 2008).reshape(-1,1)#因为X数值过大&＃xff0c;进行处理
X


y &＃61; sales
y


from sklearn.linear_model import LinearRegression#假定函数是一元二次f(x) &＃61; w1*x**2 &＃43; w2*x &＃43; b
lr &＃61; LinearRegression(fit_intercept &＃61; True)lr.fit(X,y)
#weight 权重
w &＃61; lr.coef_[0]
#bias 偏差
b &＃61; lr.intercept_
display(w,b)



plt.scatter((years-2008),sales,c &＃61; &＃39;red&＃39;,marker &＃61; &＃39;*&＃39;,s &＃61; 80)
x &＃61; np.linspace(1,12,50)
plt.plot(x,w*x &＃43; b,c &＃61; &＃39;green&＃39;)


from sklearn.linear_model import LinearRegression#假定函数是一元二次f(x) &＃61; w1*x**2 &＃43; w2*x &＃43; b
#变一元二次
lr &＃61; LinearRegression(fit_intercept &＃61; True)X2 &＃61; np.concatenate([X**2,X],axis &＃61; 1)
lr.fit(X2,y)
#weight 权重
w1,w2 &＃61; lr.coef_
#bias 偏差
b &＃61; lr.intercept_
display(w1,w2,b)


plt.scatter((years-2008),sales,c &＃61; &＃39;red&＃39;,marker &＃61; &＃39;*&＃39;,s &＃61; 80)
x &＃61; np.linspace(1,12,50)
plt.plot(x,w1*x**2 &＃43;w2*x &＃43; b,c &＃61; &＃39;green&＃39;)


f &＃61; lambda x:w1*x**2 &＃43; w2*x &＃43;b
print(&＃39;2020年销量预测&＃39;,f(12))