Automate Time Series Forecasting-using R



Hello friends,

I am trying to create a Rshiny app to automate sales forecasting.

I am using time series forecasting as the sales are monthly.

I have data from Nov 2012 to Sep 2016.

Before creating the app, I want to do manual forecasting to see which method best suits the forecasts.

I have used Holt.Winters exponential smoothing, ETS and Auto.Arima() in R.

Now when I use ETS() and Auto.Arima() (output of auto.arima is ARIMA(0,1,1) i get same value for all the forecasts.

How to interpret this? why I am getting same value?( I guess (0,1,1) is simple smoothing and hence using last value to forecast)

Also to automate forecasting what other packages/options are there in R to get accurate forecasts.

I have data for 120 SKU’s…do I need to individually check for each and every SKU…or I can use a single forecasting technique for all?