My question is pertaining to automatic forecast of multiple products. I am using a combination of 2 models to forecast my timeseries data for 190 products. The values are arranged in column format. Now the 106th, 107th and 108th product has all the values as 0. Total number of observation points are 51. The model is unable to forecast any product here on (even though they have values apart from 0s). Is it possible that the models like TBATS, StructTS, STLF have a restriction with regards to datasets having 0 values? I did read a few articles where croston has been suggested for such datasets. But mine has -ve values as well and therefore, I cannot use croston. Please suggest what canbe done?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.