I have a sales forecasting problem where I only have three months of year 2018 ( Jan, fev, march) and three months of year 2019 (oct, nov,dec) and I have to forecast the sales for all the months of 2020 (these are the only data the customer had provided to me, I cant have more). I am new with the forecasting problems and what I can conclude from these data that it is time series data but it's so limited and the gap between the two years is about 18 months so handling missing data won't be a solution for this. I have been so stack can anyone help me please. Can you suggest me what to do ?
You must use the data for the three months of 2018 in forecasting the remaining months of the same year, and the same case for the year 2019, and later you can use the data for the two years to forecast the year 2020
Bushra Abdul Hamza Abbas thank you for you answer, but how I will evaluate each model with only three months ?
( The data is on daily basis)
Filling the gap with imputation would be efficient in this case, with this limited data?
@Houda Kilani to use the date that you have to fill the missing gaps, you have to be sure that the selling are not seasonal. Otherwise it is inappropriate to forecast in that way.
- Badges
- Science topic
- Similar topics
- Thesis Research