0
$\begingroup$

I have 3 year's worth of revenue build-up data (52 weeks, weekly) from 2015 - 2018. So, it is a revenue build-up curve on a yearly basis.

Something like this, with different colors describing different years. enter image description here

Hypothetically, if I have one single data point in the future, say 4 weeks into 2019, can I do time series taking this extra data point that I have into account? Can I "join"/predict the beginning week of 2019 to the 4th week of 2019 (data pt that I have) somehow? (short of drawing a straight line to the future data pt)

Is there something like this? I'm exploring rnn, arima, prophet but none of them seems to be able to cater for this.

The only method I can think of is to use regression, but this didn't give me satisfactory results, it's under-predicting by a lot.

Looking for some ideas on how to model this.

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

You may ultimately want to predict the cumulative , but you will be better served to model the actual weekly (observed) data and predict that and then impute the annual cumulative. If you post your 4*52=208 actual values I night be able to help further.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thanks, It is cumulative revenue build-up currently. I can't really post the actual values as this is company data. $\endgroup$
    – lppier
    Sep 8, 2018 at 14:50
  • $\begingroup$ Then scale the data by subtracting a constant and then divide by another constant or simply create a similar time series to yours. $\endgroup$
    – IrishStat
    Sep 8, 2018 at 16:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.