Apriori algorithm with tags
In apriori algorithm, we can create association rules with respect to the frequencies of the corresponding data set. My question is, what if we have tags data in addition to the transaction data. For example take a transaction data set like;
| TransactionID | ItemGroups |
|---|---|
| 1 | A,C,D |
| 2 | B,C,E |
| 3 | A,B,C,E |
| 4 | B,E |
It is easy to set support, lift and confidence to reach an assosiation rule with the toy data above.
Now consider we have an additional item-tag data like this;
| Item | Tag |
|---|---|
| A | Tag1 |
| A | Tag2 |
| A | Tag3 |
| B | Tag2 |
| B | Tag3 |
| B | Tag5 |
| B | Tag7 |
| B | Tag8 |
| C | Tag7 |
| D | Tag1 |
| D | Tag8 |
| E | Tag2 |
How can it be used to improve the apriori recommendation? Another apriori with tags? But with this, since there is no unique length of the tag number wrt the item (I mean i.e. the item A has 3 tags while, B has 5), I am stuck to bridge the number of tags with the item recommendation.
Thanks in advance!
Topic association-rules recommender-system data-mining
Category Data Science