SAS EM: Association

ycenycute — Fri, 29 Oct 2021 13:06:07 GMT

I have several questions regarding association node.

1. Does it require at least one ID variable and one Target variable? Why is that? What does ID and Target do here?

2. What does maximum items mean? I can find the descriptions in the document, but still don't understand. Can someone explain? Like what is a four-way association? Why when there is sequence variable, it is reset to 2?

Maximum Items — Use the Maximum Items property to specify the maximum size of any given item set that you want to consider in an association. For the example, the default value of 4 items indicates that up to four-way associations are performed. The minimum number of Maximum Items is 1. Permissible values are nonnegative integers. If you set Maximum Items to 1 and your data source does not contain a sequence variable, then Maximum Items is reset to 2.

Re: SAS EM: Association

Amod1 — Tue, 02 Nov 2021 07:20:27 GMT

Ref: https://documentation.sas.com/doc/en/emref/14.3/n16x97j506upgin1l90wrfc1rg0l.htm

1. I think of ID variable as a customer, Target variable as Item he purchases and sequence variable as his visit number, or the date of purchase.
2. An item set contains items purchased. A frequent item set is that item set which occurs (purchased) frequently by the ID (customer). Now, maximum items tells us about the size of the item set. If 4, an item set can contain at the most 4 items.
3. 4-way associations: possible ways in which we can combine them to form any association rule. We can take 1 or 2 or 3 or 4 (all) of them in creating an association.
4. Now, in case when you have sequence (date/time) variable, you can associate 1 item (target) from 1st visit of same customer (ID) with any other item from the subsequent visits (purchase dates). That’s why we need at least 1 item in maximum items parameter. However, if we don’t have sequence variable, we are restricted to find the associations within the same visit (purchase date), as there is only one visit by that customer. In that case, association requires at least 2 items, one on LHS and the other on RHS. That’s why If you set Maximum Items to 1 and your data source does not contain a sequence variable, then Maximum Items is reset to 2.

topic SAS EM: Association in SAS Data Science

SAS EM: Association

Re: SAS EM: Association