0
$\begingroup$

Which forecasting model fits better to forecast the customers spend in the next upcoming visit? We have millions of customers, so ARIMA or any other time series modeling for each of the customers is not feasible (computationally).

Built Linear model (using a different level of moving average terms and other independent variables) and the average model (average spend per visit in the last X visits or time period). The performance of the linear model is similar to that of average model. How can I improve the linear model or build a better forecasting model?

Sample data:
Customer    Visit Date  Spend
1           1/21/2018   302
2           1/21/2018   308
3           1/23/2018   2112
4           1/35/2018   123
1           2/14/2018   234
1           2/16/2018   534
2           3/4/2018    1254
3           6/6/2018    854
1           6/2/2018    150
1           6/8/2018    720
2           9/9/2018    440
4           12/12/2018  323
3           12/15/2018  445
1           12/16/2018  660
1           as_of_today ?
2           as_of_today ?
3           as_of_today ?
4           as_of_today ?

Edit on 2019-01-31: I am not looking to forecast individually for each of the customers. I did segment the customers and built the linear models for each segment. However, want to build a better model than a linear model. So looking if there are any better forecasting techniques that can help me.

$\endgroup$
  • $\begingroup$ Do you really need individual forecast for each one of the millions of customers? For what purpose will these forecast be used? Maybe start with segmentation, time-series clustering, ... and then maybe individual forecasts for only the most important? $\endgroup$ – kjetil b halvorsen Jan 26 at 18:56
  • $\begingroup$ I am not looking to forecast individually for each of the customers. I did segment the customers and built the linear models for each segment. However, want to build a better model than a linear model. So looking if there are any better forecasting techniques that can help me. $\endgroup$ – ML101 Jan 29 at 18:33
  • $\begingroup$ Please add this new information as an edit to the OP (original post). But without more details I doubt anybody will have much to say $\endgroup$ – kjetil b halvorsen Jan 29 at 20:15
  • 1
    $\begingroup$ I'm a little confused. Are you interested in when the customer will return, how much the individual will spend, or both? $\endgroup$ – Demetri Pananos Feb 1 at 3:11
  • $\begingroup$ @DemetriPananos I am interested only in how much will the customer spend. $\endgroup$ – ML101 Feb 2 at 1:27

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.