I'm forecasting next week visit (11/21/2017-11/27/2017) based on the visit history data I have, and I am considering holidays (1 is holiday and 0 is not) in the forecasting.
I have several questions of the Aggregation and Accumulation:
1) What's the difference between Aggregation and Accumulation in the Forecasting?
2) In my example below, the hierarchy is Total----Location---Department.
For the dependent variables "Visits" I think I will choose "Sum of Values" for Aggregation and Accumulation
But I am not sure about the independent variable "Holidays", what should I choose for Aggregation and Accumulation of variable "Holidays" ? Maybe choose "Average of Values" (See details in attachments)? Any comments are welcome!
The data structure I have is below:
|Values going on||Values going on||Values going on||Values going on||Values going on|
By ACCUMULATION it is intended how the data is "aggregated" over time. For example, if you have data at daily interval, and you convert it to weekly interval, you are accumulating your data.
"Aggregation" is the aggregation over the levels of your hierarchy. "Total", is the aggregation of all the levels below.
Typically you choose aggregation=accumulation=total for the dependent variable, and aggregation=accumulation=average for the independent variable, but it is the context that dictates what to use. In the case of holidays I would chose average.
Are you using Forecast Studio? I suppose you are aware that Forecast Server can generate the holiday variables for you.
1) In short:
>> Aggregation is aggregating along the hierarchy, i.e., from low level to high level in the hierarchy
>> Accumulation refers to accumulate by time domain, i.e., from higher frequency (daily) to lower frequency (weekly). It is not dependent on whether you have a hierarchy
In details, accumulation can be either of the following:
• the process of converting a time series that has no fixed interval into a time series that has a fixed interval (such as hourly or monthly)
• the process of converting a time series that has a fixed interval into a time series with a lower frequency time interval (such as hourly into daily)
2) For Holidays, if you treat it as a dummy variable, I suggest you use MIN, MAX or N to do aggregation or accumulation.
Registration is open! SAS is returning to Vegas for an AI and analytics experience like no other! Whether you're an executive, manager, end user or SAS partner, SAS Innovate is designed for everyone on your team. Register for just $495 by 12/31/2023.
If you are interested in speaking, there is still time to submit a session idea. More details are posted on the website.
Learn how to run multiple linear regression models with and without interactions, presented by SAS user Alex Chaplin.
Find more tutorials on the SAS Users YouTube channel.