拓端数据tecdat|R语言神经网络模型预测车辆数量时间序列

1.  
2.  
## Date Actual
3.  
##  
4.  
## 1 2016-01-11 26
5.  
## 2 2016-01-18 27
6.  
## 3 2016-01-25 28
7.  
## 4 2016-02-01 22
8.  
## 5 2016-02-08 27
9.  
## 6 2016-02-15 31
 
1.  
dat%>%ggplot(aes(Date,Actual))+
2.  
labs(title = "耐久性车辆预测", x = "时间", y = "耐用车辆数量",
3.  
subtitle = " 2016 & 2017年数据") +

1.  
#数据必须为数字，tsclean函数才能正常工作
2.  
dat%>%dplyr::rename(Date=Week)%>%na.omit()
3.  
dat_ts = ts(dat["实际数量"])
 
nn(data, lambda=0.5)1.   
2.   
3.  
Average of 20 networks, each of which is
4.  
a 1-1-1 network with 4 weights
5.  
options were - linear output units 
6.   
7.  
sigma^2 estimated as 0.2705

plot(dat_ts) + autolayer(sim)

plot(fcast)