Оршил
1. Багц болон өгөгдөл
library(tidyverse)
library(hexView)
library(uroot)
library(forecast)
library(xtable)
library(knitr)
data<-readEViews("/Users/Mongol/Documents/Eviews/ElementsOfForecasting/fcst_06/fcst6input.wf1")
mydata<-filter(data,between(Date,as.Date("1946-01-01"),as.Date("1994-11-01")))
house.starts<-ts(mydata$HSTARTS,start = c(1946,1),end = c(1994,11),frequency = 12)
house_starts<-head(house.starts,576)
house_starts2<-tail(house.starts,59)
# Figure 6.4 Housing starts, 1946.01-1994.11
plot.ts(house_starts,ylab="Starts",main="Figure 6.4 Housing Starts 1946.01-1994.11",bty="l"
,ylim=c(0,250),yaxt="n",xaxt="n",type="l",lwd=2)
axis(2,at=c(seq(0,250,50)),labels = c(seq(0,250,50)),las=1)
axis(1,at=c(seq(1950,1990,5)),labels = c(seq(50,90,5)))
2. Улирлын нөлөө
# Figure 6.5 Housing starts, 1990.01-1994.11
plot(house_starts2,ylab="Starts",yaxt="n",type="l",lwd=2
,main="Figure 6.5 Housing starts, 1990.01-1994.11",bty="l",ylim=c(40,160))
# axis(1,at=c(seq(as.Date("1990/1/1"),as.Date("1994/11/1"),by = "2 quarter"))
# ,labels = c("90:01","90:07","91:01","91:07","92:01","92:07","93:01","93:07","94:01","94:07"))
axis(2,at=c(seq(40,160,20)),labels = c(seq(40,160,20)),las=1)
D<-seasonal.dummies(house_starts)
reg1<-lm(house_starts~D-1)
result1<-summary(reg1)
result1$coefficients## Estimate Std. Error t value Pr(>|t|)
## DSD1 86.50417 4.029055 21.47009 3.463337e-75
## DSD2 89.50417 4.029055 22.21468 5.073476e-79
## DSD3 122.88333 4.029055 30.49929 2.020417e-121
## DSD4 142.16875 4.029055 35.28588 7.304846e-145
## DSD5 147.50000 4.029055 36.60908 3.821490e-151
## DSD6 145.99792 4.029055 36.23627 2.201830e-149
## DSD7 139.11250 4.029055 34.52733 3.200033e-141
## DSD8 138.41667 4.029055 34.35462 2.179227e-140
## DSD9 130.56250 4.029055 32.40524 6.972702e-131
## DSD10 134.09167 4.029055 33.28117 3.546630e-135
## DSD11 111.83333 4.029055 27.75671 1.484717e-107
## DSD12 92.15833 4.029055 22.87344 2.024047e-82ger<-function(regression,result,dependent.var){print(data.frame(value=c(result$r.squared,result$adj.r.squared,result$sigma,deviance(regression),logLik(regression),sum(diff(regression$residuals,lag = 1)^2)/sum(regression$residuals^2),mean(dependent.var),sd(dependent.var),AIC(regression),BIC(regression),result$r.squared/(1-result$r.squared)*(length(dependent.var)-(length(regression$coefficients))/(length(regression$coefficients)-1)),pf(result$r.squared/(1-result$r.squared)*(length(dependent.var)-(length(regression$coefficients))/(length(regression$coefficients)-1)),length(regression$coefficients)-1,length(dependent.var)-(length(regression$coefficients)),lower.tail = F)),row.names = c("Rsquared","Adjusted Rsquared","SE","SSR","Log lokelihood","Durbin-Watson stat.","Mean dependent var.","SD dependent var.","Akaike info criterion","Schwarz criterion","F-statistic","p-value")))}
ger(reg1,result1,house_starts)## value
## Rsquared 9.536595e-01
## Adjusted Rsquared 9.526736e-01
## SE 2.791411e+01
## SSR 4.394675e+05
## Log lokelihood -2.728825e+03
## Durbin-Watson stat. 1.541402e-01
## Mean dependent var. 1.233944e+02
## SD dependent var. 3.521775e+01
## Akaike info criterion 5.483651e+03
## Schwarz criterion 5.540280e+03
## F-statistic 1.183129e+04
## p-value 0.000000e+00| Estimate | Std. Error | t value | Pr(>|t|) | |
|---|---|---|---|---|
| DSD1 | 86.50417 | 4.029055 | 21.4701 | 0 |
| DSD2 | 89.50417 | 4.029055 | 22.2147 | 0 |
| DSD3 | 122.88333 | 4.029055 | 30.4993 | 0 |
| DSD4 | 142.16875 | 4.029055 | 35.2859 | 0 |
| DSD5 | 147.50000 | 4.029055 | 36.6091 | 0 |
| DSD6 | 145.99792 | 4.029055 | 36.2363 | 0 |
| DSD7 | 139.11250 | 4.029055 | 34.5273 | 0 |
| DSD8 | 138.41667 | 4.029055 | 34.3546 | 0 |
| DSD9 | 130.56250 | 4.029055 | 32.4052 | 0 |
| DSD10 | 134.09167 | 4.029055 | 33.2812 | 0 |
| DSD11 | 111.83333 | 4.029055 | 27.7567 | 0 |
| DSD12 | 92.15833 | 4.029055 | 22.8734 | 0 |
# Figure 6.6 Residual plot
par(mar=c(4,4,4,4))
plot.ts(house_starts,type="l",lwd=2,main="Figure 6.6 Residual plot",ylab="",yaxt="n",xaxt="n"
,ylim=c(-150,250),bty="o")
lines(ts(reg1$fitted.values,start = c(1946,1),end = c(1993,12),frequency = 12)
,lwd=2,lty=2,col="blue")
axis(4,at=c(seq(0,250,50)),labels =c(seq(0,250,50)),las=2)
axis(1,at=c(seq(1950,1990,5)),labels = c(seq(50,90,5)),cex.axis=1,xlab="")
par(new=T)
resid1<-ts(reg1$residuals,start = c(1946,1),end = c(1993,12),frequency = 12)
plot(resid1,col="red",lwd=2,type="l",ylab="",yaxt="n",ylim=c(-100,200),xaxt="n",xlab="")
axis(2,at=c(seq(-100,100,50)),labels = c(seq(-100,100,50)),las=2,cex.axis=1)
abline(h=mean(resid1))
abline(h=mean(resid1)+result1$sigma)
abline(h=mean(resid1)-result1$sigma)
legend("topleft",c("Residual","Actual","Fitted"),lty=c(1,1,2),
col=c("red","black","blue"),cex=0.7)
# Figure 6.7 Estimated seasonal factors, housing starts
plot(reg1$coefficients,main="Figure 6.7 Estimated seasonal factors, housing starts",type="l"
,lwd=2,yaxt="n",xlab="Season",ylab="Seasonal factors",ylim=c(80,160),bty="l")
axis(2,at=c(seq(80,160,20)),labels = c(seq(80,160,20)),las=2)
3. Таамаглал
house.forecast<-forecast(tail(house_starts,48),h=11,level = .95)
plot(house.forecast,xlim=c(1990,1995),type="l",lwd=2,ylim = c(0,250)
,ylab="History forecast",xlab = "Time",yaxt="n"
,main="Figure 6.8 Housing starts:History and forecast")
axis(2,at=c(seq(0,250,50)),labels = c(seq(0,250,50)),las=2)
# Eviews plot
plot(tail(house_starts,48),xlim=c(1990,1995),type="l",lwd=2,ylim = c(0,250)
,ylab="History forecast",xlab = "Time",yaxt="n"
,main="Figure 6.8 Housing starts:History and forecast")
axis(2,at=c(seq(0,250,50)),labels = c(seq(0,250,50)),las=2)
lines(house.forecast$mean,lwd=2,lty=2,col="black")
lines(house.forecast$upper,lty=2,col="black")
lines(house.forecast$lower,lty=2,col="black")
# Figure 6.9 Housing starts: History, 1990.01-1993.12; and forecast and realization, 1994.01-1994.11
plot(house_starts2,xlim=c(1990,1995),type="l",lwd=2,ylim = c(0,250)
,ylab="History forecast",xlab = "Time",yaxt="n"
,main="Figure 6.9 Housing starts:History and forecast")
axis(2,at=c(seq(0,250,50)),labels = c(seq(0,250,50)),las=2)
lines(house.forecast$mean,lwd=2,col="blue")
lines(house.forecast$upper,lwd=2,col="gray")
lines(house.forecast$lower,lwd=2,col="gray")
4. Ашигласан материал:
Diebold https://www.sas.upenn.edu/~fdiebold/Textbooks.html
Хичээлтэй холбоотой санал сэтгэгдлийг мэйл хаяг болон фэйсбүүк хаягаар нээлттэй хүлээн авах болно .