Am trying to forecast using time series method called ARIMA. I have followed steps to build a time series model displayed in the code below. My challenge is on (Merging Actual and Forecast in One Series) and Remove Transformation from Series.
salesData<- read.csv("C:/Users/sales.csv") #converting data into time series salesData<-ts(salesData,start = 2010, frequency = 4) # Checking for stationarity library(tseries) adf.test(salesData) # transforming data to stationary ndiffs(salesData) diff_series <- diff(salesData) tsdisplay(diff_series) ## Augmented Dickey-Fuller Test for stationarity adf.test(salesData) # Arima forecasting auto generated, 1-step ahead arimafore = forecast(auto.arima(salesData), h = 1) summary(arimafore) autoplot(arimafore,xlab="years", ylab="sales") #Final forecasting after removing difference(transformation) ?????????
My question is how do i best Merge Actual and Forecast in One Series and then'transform' the forecast data 'back' so I can plot it along with the original time-series using Differencing method ?