R：将年度预测功能转换为每月当量以进行时间序列数据分析

问题描述 投票：0回答：1

我有一个用于年度时间序列预测的函数；现在，我想在现有函数的基础上制作一个类似的函数来预测每月时间序列数据。

现有功能

library(forecast)
library(zoo)

set.seed(289805)
y <- arima.sim(n = 10, model = list(ar = 0.8, order = c(1, 0, 0)), sd = 1)
y <- ts(y, frequency = 1, start = c(1981))


#-- Extract Training Data, Fit the Wrong Model, and Forecast
yt <- window(y, end = 1990)

#yfit <- Arima(yt,order = c(3, 1, 4))
yfit <- forecast::auto.arima(yt)

yfor <- forecast(yfit)

#---Extract the Data for ggplot using funggcast()
funggcast <- function(dn, fcast){

  en <- max(time(fcast$mean)) # Extract the max date used in the forecast

  # Extract Source and Training Data
  ds <- as.data.frame(window(dn, end = en))
  names(ds) <- 'observed'
  ds$Year <- as.Date(time(window(dn, end = en)))

  # Extract the Fitted Values (need to figure out how to grab confidence intervals)
  dfit <- as.data.frame(fcast$fitted)
  dfit$Year <- as.Date(time(fcast$fitted))
  names(dfit)[1] <- 'fitted'

  ds <- merge(ds, dfit, all.x = T) # Merge fitted values with source and training data

  # Extract the Forecast values and confidence intervals
  dfcastn <- as.data.frame(fcast)
  dfcastn$Year <- as.Date(paste(row.names(dfcastn), "01", "01", sep = "-"))
  names(dfcastn) <- c('forecast', 'lo80', 'hi80', 'lo95', 'hi95', 'Year')

  pd <- merge(ds, dfcastn, all = T) # final data.frame for use in ggplot
  return(pd)

}

pd <- funggcast(y ,yfor)

pd
#         Year   observed     fitted  forecast       lo80     hi80      lo95     hi95
#1  1981-01-01 -0.2505125 -0.1002050        NA         NA       NA        NA       NA
#2  1982-01-01  0.8855338 -0.2004100        NA         NA       NA        NA       NA
#3  1983-01-01  0.1975730  0.7084271        NA         NA       NA        NA       NA
#4  1984-01-01  1.5355101  0.1580584        NA         NA       NA        NA       NA
#5  1985-01-01  4.6956397  1.2284081        NA         NA       NA        NA       NA
#6  1986-01-01  4.4414254  3.7565120        NA         NA       NA        NA       NA
#7  1987-01-01  3.0230805  3.5531406        NA         NA       NA        NA       NA
#8  1988-01-01  1.8058732  2.4184646        NA         NA       NA        NA       NA
#9  1989-01-01  1.1661036  1.4446987        NA         NA       NA        NA       NA
#10 1990-01-01  1.1631486  0.9328829        NA         NA       NA        NA       NA
#11 1991-01-01         NA         NA 0.9305189 -0.8119452 2.672983 -1.734350 3.595388
#12 1992-01-01         NA         NA 0.7444152 -1.4870278 2.975858 -2.668282 4.157113
#13 1993-01-01         NA         NA 0.5955322 -1.8990535 3.090118 -3.219607 4.410672
#14 1994-01-01         NA         NA 0.4764258 -2.1728901 3.125742 -3.575353 4.528205
#15 1995-01-01         NA         NA 0.3811406 -2.3626271 3.124908 -3.815090 4.577371
#16 1996-01-01         NA         NA 0.3049125 -2.4976344 3.107459 -3.981213 4.591038
#17 1997-01-01         NA         NA 0.2439300 -2.5955971 3.083457 -4.098752 4.586612
#18 1998-01-01         NA         NA 0.1951440 -2.6677997 3.058088 -4.183351 4.573639
#19 1999-01-01         NA         NA 0.1561152 -2.7217152 3.033946 -4.245147 4.557377
#20 2000-01-01         NA         NA 0.1248922 -2.7624254 3.012210 -4.290879 4.540663

我尝试过的

library(forecast)
library(zoo)

set.seed(289805)
y <- arima.sim(n = 10, model = list(ar = 0.8, order = c(1, 0, 0)), sd = 1)
y <- ts(y, frequency = 12, start = c(1981, 1))


#-- Extract Training Data, Fit the Wrong Model, and Forecast
yt <- window(y, end = c(1981, 10))

#yfit <- Arima(yt,order = c(3, 1, 4))
yfit <- forecast::auto.arima(yt)

yfor <- forecast(yfit)

#---Extract the Data for ggplot using funggcast()
funggcast <- function(dn, fcast){

  en <- max(time(fcast$mean)) # Extract the max date used in the forecast

  # Extract Source and Training Data
  ds <- as.data.frame(window(dn, end = en))
  names(ds) <- 'observed'
  ds$Year <- as.Date(time(window(dn, end = en)))

  # Extract the Fitted Values (need to figure out how to grab confidence intervals)
  dfit <- as.data.frame(fcast$fitted)
  dfit$Year <- as.Date(time(fcast$fitted))
  names(dfit)[1] <- 'fitted'

  ds <- merge(ds, dfit, all.x = T) # Merge fitted values with source and training data

  # Extract the Forecast values and confidence intervals
  dfcastn <- as.data.frame(fcast)
  dfcastn$Year <- as.Date(paste(row.names(dfcastn), "01", "01", sep = "-"))
  names(dfcastn) <- c('forecast', 'lo80', 'hi80', 'lo95', 'hi95', 'Year')

  pd <- merge(ds, dfcastn, all = T) # final data.frame for use in ggplot
  return(pd)

}

pd <- funggcast(y ,yfor)

pd

我收到此错误消息：

charToDate(x) 中的错误： 字符串不是标准的明确格式 另外： 警告消息： 1：在 window.default(x, ...) 中：“结束”值未更改 2：在 window.default(x, ...) 中：'end' 值未更改

我想要什么

我需要 

Year
，而不是
Month/Year

列

我想要这样的东西

data frame
，其中提供的时间序列数据是从1981年1月1日到1982年2月1日到1981年10月1日的每月数据。样本外预测将是从11月开始的每月时间序列预测1981年1月1日至1981年12月1日至1982年8月1日，如下所示：

#    Year/Month   observed     fitted  forecast       lo80     hi80      lo95     hi95
#1  1981-01-01 -0.2505125 -0.1002050        NA         NA       NA        NA       NA
#2  1981-02-01  0.8855338 -0.2004100        NA         NA       NA        NA       NA
#3  1981-03-01  0.1975730  0.7084271        NA         NA       NA        NA       NA
#4  1981-04-01  1.5355101  0.1580584        NA         NA       NA        NA       NA
#5  1981-05-01  4.6956397  1.2284081        NA         NA       NA        NA       NA
#6  1981-06-01  4.4414254  3.7565120        NA         NA       NA        NA       NA
#7  1981-07-01  3.0230805  3.5531406        NA         NA       NA        NA       NA
#8  1981-08-01  1.8058732  2.4184646        NA         NA       NA        NA       NA
#9  1981-09-01  1.1661036  1.4446987        NA         NA       NA        NA       NA
#10 1991-10-01  1.1631486  0.9328829        NA         NA       NA        NA       NA
#11 1981-11-01         NA         NA 0.9305189 -0.8119452 2.672983 -1.734350 3.595388
#12 1981-12-01         NA         NA 0.7444152 -1.4870278 2.975858 -2.668282 4.157113
#13 1982-01-01         NA         NA 0.5955322 -1.8990535 3.090118 -3.219607 4.410672
#14 1982-02-01         NA         NA 0.4764258 -2.1728901 3.125742 -3.575353 4.528205
#15 1982-03-01         NA         NA 0.3811406 -2.3626271 3.124908 -3.815090 4.577371
#16 1982-04-01         NA         NA 0.3049125 -2.4976344 3.107459 -3.981213 4.591038
#17 1982-05-01         NA         NA 0.2439300 -2.5955971 3.083457 -4.098752 4.586612
#18 1982-06-01         NA         NA 0.1951440 -2.6677997 3.058088 -4.183351 4.573639
#19 1982-07-01         NA         NA 0.1561152 -2.7217152 3.033946 -4.245147 4.557377
#20 1982-08-01         NA         NA 0.1248922 -2.7624254 3.012210 -4.290879 4.540663
r dataframe time-series forecasting forecast
1个回答
0
投票

问题是

dfcastn
的行名称已经包含月份名称的缩写。为了解决这个问题，您可以使用例如
as.Date(paste("01", row.names(dfcastn)), format = "%d %B %Y")
转换为正确的日期。

注意：由于我在德语区域设置上运行，因此我必须暂时切换到英语区域设置，以将月份名称的英语缩写转换为正确的日期。如果您的情况并非如此，您可以删除相应的代码行。

library(forecast)
library(zoo)

funggcast <- function(dn, fcast) {
  en <- max(time(fcast$mean)) # Extract the max date used in the forecast

  # Extract Source and Training Data
  ds <- as.data.frame(window(dn, end = en))
  names(ds) <- "observed"
  ds$Year <- as.Date(time(window(dn, end = en)))

  # Extract the Fitted Values (need to figure out how to grab confidence intervals)
  dfit <- as.data.frame(fcast$fitted)
  dfit$Year <- as.Date(time(fcast$fitted))
  names(dfit)[1] <- "fitted"

  ds <- merge(ds, dfit, all.x = T) # Merge fitted values with source and training data

  # Extract the Forecast values and confidence intervals
  dfcastn <- as.data.frame(fcast)
  
  # Switch to english locale
  old <- Sys.getlocale(category = "LC_TIME")
  Sys.setlocale(category = "LC_TIME", locale = "en_US.UTF-8")
  dfcastn$Year <- as.Date(paste("01", row.names(dfcastn)), format = "%d %B %Y")
  # Restore old locale
  Sys.setlocale(category = "LC_TIME", locale = old)
  
  names(dfcastn) <- c("forecast", "lo80", "hi80", "lo95", "hi95", "Year")
  
  
  merge(ds, dfcastn, all = T) # final data.frame for use in ggplot
}

pd <- funggcast(y, yfor)
#> Warning in window.default(x, ...): 'end' value not changed

#> Warning in window.default(x, ...): 'end' value not changed

pd
#>          Year   observed     fitted    forecast       lo80     hi80      lo95
#> 1  1981-01-01 -0.2505125 -0.1002050          NA         NA       NA        NA
#> 2  1981-02-01  0.8855338 -0.2004100          NA         NA       NA        NA
#> 3  1981-03-01  0.1975730  0.7084271          NA         NA       NA        NA
#> 4  1981-04-01  1.5355101  0.1580584          NA         NA       NA        NA
#> 5  1981-05-01  4.6956397  1.2284081          NA         NA       NA        NA
#> 6  1981-06-01  4.4414254  3.7565120          NA         NA       NA        NA
#> 7  1981-07-01  3.0230805  3.5531406          NA         NA       NA        NA
#> 8  1981-08-01  1.8058732  2.4184646          NA         NA       NA        NA
#> 9  1981-09-01  1.1661036  1.4446987          NA         NA       NA        NA
#> 10 1981-10-01  1.1631486  0.9328829          NA         NA       NA        NA
#> 11 1981-11-01         NA         NA 0.930518905 -0.8119452 2.672983 -1.734350
#> 12 1981-12-01         NA         NA 0.744415169 -1.4870278 2.975858 -2.668282
#> 13 1982-01-01         NA         NA 0.595532171 -1.8990535 3.090118 -3.219607
#> 14 1982-02-01         NA         NA 0.476425765 -2.1728901 3.125742 -3.575353
#> 15 1982-03-01         NA         NA 0.381140635 -2.3626271 3.124908 -3.815090
#> 16 1982-04-01         NA         NA 0.304912527 -2.4976344 3.107459 -3.981213
#> 17 1982-05-01         NA         NA 0.243930036 -2.5955971 3.083457 -4.098752
#> 18 1982-06-01         NA         NA 0.195144041 -2.6677997 3.058088 -4.183351
#> 19 1982-07-01         NA         NA 0.156115242 -2.7217152 3.033946 -4.245147
#> 20 1982-08-01         NA         NA 0.124892201 -2.7624254 3.012210 -4.290879
#> 21 1982-09-01         NA         NA 0.099913767 -2.7934593 2.993287 -4.325119
#> 22 1982-10-01         NA         NA 0.079931018 -2.8173109 2.977173 -4.351018
#> 23 1982-11-01         NA         NA 0.063944818 -2.8357705 2.963660 -4.370787
#> 24 1982-12-01         NA         NA 0.051155858 -2.8501413 2.952453 -4.385995
#> 25 1983-01-01         NA         NA 0.040924689 -2.8613844 2.943234 -4.397774
#> 26 1983-02-01         NA         NA 0.032739753 -2.8702168 2.935696 -4.406949
#> 27 1983-03-01         NA         NA 0.026191804 -2.8771790 2.929563 -4.414131
#> 28 1983-04-01         NA         NA 0.020953444 -2.8826825 2.924589 -4.419775
#> 29 1983-05-01         NA         NA 0.016762757 -2.8870429 2.920568 -4.424225
#> 30 1983-06-01         NA         NA 0.013410206 -2.8905040 2.917324 -4.427743
#> 31 1983-07-01         NA         NA 0.010728165 -2.8932555 2.914712 -4.430532
#> 32 1983-08-01         NA         NA 0.008582533 -2.8954456 2.912611 -4.432745
#> 33 1983-09-01         NA         NA 0.006866027 -2.8971906 2.910923 -4.434505
#> 34 1983-10-01         NA         NA 0.005492822 -2.8985820 2.909568 -4.435906
#>        hi95
#> 1        NA
#> 2        NA
#> 3        NA
#> 4        NA
#> 5        NA
#> 6        NA
#> 7        NA
#> 8        NA
#> 9        NA
#> 10       NA
#> 11 3.595388
#> 12 4.157113
#> 13 4.410672
#> 14 4.528205
#> 15 4.577371
#> 16 4.591038
#> 17 4.586612
#> 18 4.573639
#> 19 4.557377
#> 20 4.540663
#> 21 4.524946
#> 22 4.510880
#> 23 4.498677
#> 24 4.488307
#> 25 4.479623
#> 26 4.472429
#> 27 4.466514
#> 28 4.461681
#> 29 4.457750
#> 30 4.454564
#> 31 4.451988
#> 32 4.449910
#> 33 4.448237
#> 34 4.446892
最新问题
© www.soinside.com 2019 - 2024. All rights reserved.