├── README.md
└── TimeSeriesExample.R


/README.md:
--------------------------------------------------------------------------------
1 | # Master


--------------------------------------------------------------------------------
/TimeSeriesExample.R:
--------------------------------------------------------------------------------
 1 | #Load Libraries
 2 | library("fUnitRoots")
 3 | library(lmtest)
 4 | library("forecast")
 5 | library(FitAR)
 6 | 
 7 | #import data
 8 | data<-read.csv("data.csv",header=TRUE)
 9 | #convert to time series
10 | tsData<- ts(data[2:35,2],start=c(2011,1),frequency=12)
11 | plot(tsData)
12 | #decompose into time series components
13 | timeseriescomponents <- decompose(tsData)
14 | plot(timeseriescomponents)
15 | #detemine stationarity of data
16 | urkpssTest(tsData, type = c("tau"), lags = c("short"),use.lag = NULL, doplot = TRUE)
17 | tsstationary<-diff(tsData, differences=1)
18 | plot(tsstationary)
19 | acf(tsData,lag.max=34)
20 | #remove seasonality
21 | timeseriesseasonallyadjusted <- tsData- timeseriescomponents$seasonal
22 | plot(timeseriesseasonallyadjusted)
23 | tsstationary <- diff(timeseriesseasonallyadjusted, differences=1)
24 | plot(tsstationary)
25 | par(mfrow=c(2,1))
26 | acf(tsstationary, lag.max=34) 
27 | pacf(tsstationary, lag.max=34)
28 | #fit the model
29 | fitARIMA<-arima(tsData, order=c(1,1,1),seasonal = list(order = c(1,0,0), period = 12),method="ML")
30 | fitARIMA
31 | #significance of coefficients
32 | coeftest(fitARIMA)
33 | par(mfrow=c(1,1))
34 | acf(fitARIMA$residuals)
35 | #residual diagnostics
36 | boxresult<-LjungBoxTest (fitARIMA$residuals,k=2,StartLag=1) # residual?? or the original series?
37 | par(mfrow=c(2,1))
38 | plot(boxresult[,3],main="Ljung-Box Q Test", ylab="P-values", xlab="Lag")
39 | qqnorm(fitARIMA$residuals)
40 | qqline(fitARIMA$residuals)
41 | 
42 | auto.arima(tsData, trace=TRUE)
43 | 
44 | #forcast future values
45 | par(mfrow=c(1,1))
46 | predict(fitARIMA,n.ahead = 5)
47 | futurVal <- forecast.Arima(fitARIMA,h=10, level=c(99.5))
48 | plot.forecast(futurVal)
49 | 
50 | 


--------------------------------------------------------------------------------