r语言 - 绘制拟合的面板数据帧与实际作为时间序列

Question

我想要求 ho 将日期变量包含在每个面板数据模型估计器的所有实际值和拟合值的数据帧中。

使用以下代码时：

library(plm)
library(dplyr)
YEAR <- c(2015, 2016, 2017, 2018,
2015, 2016, 2017, 2018,
2015, 2016, 2017, 2018,
2015, 2016, 2017, 2018)
YEAR <- as.Date(YEAR)
UNIT <- c("A", "A", "A", "A",
"B", "B", "B", "B", 
"C", "C", "C", "C",
"D", "D", "D", "D")
Y <- sample(100:1000, 16)
X1 <- sample(10:50, 16)
X2 <- sample(20:60, 16)
data <- data.frame(YEAR, UNIT, Y)
crime.p <- pdata.frame(data,index=c("UNIT","YEAR")) 
fixedeff <- plm(log(Y)~X1 + X2,data=crime.p,model="within")
randomeff <- plm(log(Y)~X1  +X2,data=crime.p,model="random") 
firstdiff <- plm(log(Y)~X1 + X2,data=crime.p,model="fd") 
fixx_results <- data.frame(fitted = predict(fixedeff), 
residuals = fixedeff$residuals ) %>% 
mutate( actual = fitted + residuals, model = "fixed" ) %>% 
select(actual, model, fitted, residuals) 
random_results <- data.frame( fitted = predict(randomeff), residuals = randomeff$residuals ) %>%
mutate( actual = fitted + residuals, model = "random" ) %>% 
select(actual, model, fitted, residuals) 
fd_results <- data.frame( fitted = predict(firstdiff ), residuals = firstdiff $residuals ) %>%
mutate( actual = fitted + residuals, model = "fd" ) %>%
select(actual, model, fitted, residuals) 
fitted_res_all <- rbind(fixx_results, random_results, fd_results)
fitted_res_all

fitted_res_all数据帧不包含日期变量，我不知道如何将其添加到其中。我知道它可以"手工"完成，但我希望它尽可能通用，以便将来可以重复使用。

问题还在于每个面板数据估计器都有不同数量的观测值。 我最终想得到的是实际和拟合的时间序列图，但在 X 轴上有可变日期，在这个数据集中我们有早期数据，但有时他会每月有

Answer 1

将日期添加到数据帧应该可以解决您的问题

fixx_results <- data.frame(fitted = predict(fixedeff), 
residuals = fixedeff$residuals, date = rownames(crime.p)) %>% 
mutate( actual = fitted + residuals, model = "fixed" ) %>% 
select(actual, model, fitted, residuals, date) 
random_results <- data.frame( fitted = predict(randomeff), residuals = randomeff$residuals, date = rownames(crime.p)) %>%
mutate( actual = fitted + residuals, model = "random" ) %>% 
select(actual, model, fitted, residuals, date) 
fd_results <- data.frame( fitted = predict(firstdiff), residuals = firstdiff $residuals, date = rownames(crime.p)[4:length(crime.p)] ) %>%
mutate( actual = fitted + residuals, model = "fd" ) %>%
select(actual, model, fitted, residuals, date) ```