I have a year worth of weekly data and I need to predict end of year sales and beginning 2 weeks of next year sales in R.
| Date | Sales |
|---|---|
| 1/1/2020 | 2000 |
| 1/8/2020 | 2004 |
| 1/15/2020 | 3001 |
| 1/22/2020 | 2004 |
| 1/29/2020 | 4002 |
| 2/5/2020 | 2003 |
| 2/12/2020 | 5000 |
| 2/19/2020 | 2222 |
| 2/26/2020 | 3300 |
| 3/4/2020 | 6000 |
| 3/11/2020 | 6600 |
| 3/18/2020 | 4003 |
| 3/25/2020 | 5004 |
| 4/1/2020 | 4003 |
| 4/8/2020 | 7003 |
| 4/15/2020 | 7323 |
| 4/22/2020 | 4566 |
| 4/29/2020 | 8000 |
| 5/6/2020 | 8001 |
| 5/13/2020 | 6002 |
| 5/20/2020 | 2001 |
| 5/27/2020 | 4002 |
| 6/3/2020 | 2001 |
| 6/10/2020 | 4003 |
| 6/17/2020 | 8000 |
| 6/24/2020 | 9002 |
| 7/1/2020 | 9002 |
| 7/8/2020 | 8009 |
| 7/15/2020 | 9004 |
| 7/22/2020 | 4003 |
| 7/29/2020 | 1000 |
| 8/5/2020 | 2009 |
| 8/12/2020 | 3099 |
| 8/19/2020 | 4009 |
| 8/26/2020 | 4008 |
| 9/2/2020 | 4001 |
| 9/9/2020 | 2008 |
| 9/16/2020 | 8004 |
| 9/23/2020 | 8003 |
| 9/30/2020 | 8054 |
| 10/7/2020 | 8043 |
| 10/14/2020 | 9087 |
| 10/21/2020 | 9034 |
| 10/28/2020 | 9021 |
| 11/4/2020 | 6004 |
| 11/11/2020 | 6000 |
| 11/18/2020 | 6006 |
| 11/25/2020 | 4009 |
| 12/2/2020 | 4568 |
| 12/9/2020 | 8009 |
| 12/16/2020 | 8006 |
| 12/23/2020 | 8654 |
| 12/30/2020 | 5680 |
- I need to take till October month and predict November and December sales
- I need to predict for next two quarters
- I need to predict weekly for end of 2020 and beginning of 2021 January
I used
past<- window(td, end = c(2020,50))
fit <- tbats(td)
fc <- forecast(fit, h=10)
model1<- HoltWinters(past,seasonal = "additive")
pred<- predict(model1,n.ahead = 10,prediction.interval = T)
but couldnt get the result, either the prediction goes to 2022 or beyond that. any idea how to get this and also suggest whether to use any new method to predict.