Identifying product items based on surge activity

1. A method comprising:

using a machine learning model to identify a first item of interest based on an activity in a marketplace system, the machine learning model corresponding to one or more machine learning layers;

determining a first set of user accounts associated with the first item of interest;

determining a second set of user accounts associated with listing the first item of interest for sale;

using the one or more machine learning layers to determine an item affinity pattern between the first set of user accounts and the second set of user accounts, the determining the item affinity pattern being based on a first set of items associated with the first set of user accounts and a second set of items listed by the second set of user accounts;

determining a third set of user accounts based on the first set of user accounts, the second set of user accounts, and the item affinity pattern; and

causing display of a notification on one or more client devices associated with the third set of user accounts, the notification indicating an opportunity to list the first item of interest for sale.

2. The method of claim 1 , wherein the activity includes a surge in sales of the first item of interest in the marketplace system, the using the machine learning model to identify the first item of interest based on an activity in a marketplace system further comprises:

determining a plurality of activity counts associated with a plurality of items based on the activity;

normalizing the plurality of activity counts into a plurality of confidence scores;

determining one or more confidence scores that are no less than a minimum confidence threshold; and

identifying the first item of interest based on the one or more confidence scores.

3. The method of claim 1 , wherein the activity includes a surge in listings of the first item of interest for sale in the marketplace system.

4. The method of claim 1 , wherein the activity is identified by the machine learning model.

5. The method of claim 1 , wherein the determining the item affinity pattern further comprises:

determining a third set of items that are common to both of the first set of items purchased by the first set of user accounts and the second set of items listed by the second set of user accounts;

determining a second item of interest from the third set of items; and

determining the third set of user accounts associated with listing the second item of interest in the marketplace system.

6. The method of claim 5 , wherein the first set of items excludes the first item of interest, and the second set of items excludes the first item of interest.

7. The method of claim 5 , wherein the determining the item affinity pattern further comprises:

determining a first ranking for the first set of items;

determining a second ranking for the second set of items; and

determining the second item of interest in the third set of items based on the first ranking and the second ranking.

8. The method of claim 7 , wherein the determining the second item of interest in the third set of items based on the first ranking and the second ranking comprises:

determining, in the third set of items, a particular item having a highest ranking based on the first ranking and the second ranking.

9. The method of claim 1 , further comprising:

determining a subset of user accounts of the second set of user accounts, the subset of user accounts not being associated with a current listing of the first item of interest for sale; and

causing the notification to be sent to the subset of user accounts of the second set of user accounts.

10. A system comprising:

one or more processors; and

a computer-storage medium having instructions stored there on that, when executed by the one or more processors, cause the system to perform operations comprising:

using a machine learning model to identify a first item of interest based on an activity in a marketplace system, the machine learning model corresponding to one or more machine learning layers;

determining a first set of user accounts associated with the first item of interest;

determining a second set of user accounts associated with listing the first item of interest for sale;

using the one or more machine learning layers to determine an item affinity pattern between the first set of user accounts and the second set of user accounts, the determining the item affinity pattern being based on a first set of items associated with the first set of user accounts and a second set of items listed by the second set of user accounts;

determining a third set of user accounts based on the first set of user accounts, the second set of user accounts, and the item affinity pattern; and

causing display of a notification on one or more client devices associated with the third set of user accounts, the notification indicating an opportunity to list the first item of interest for sale.

11. The system of claim 10 , wherein the activity includes a surge in sales of the first item of interest in the marketplace system, the using the machine learning model to identify the first item of interest based on an activity in a marketplace system further comprises:

determining a plurality of activity counts associated with a plurality of items based on the activity;

normalizing the plurality of activity counts into a plurality of confidence scores;

determining one or more confidence scores that are no less than a minimum confidence threshold; and

identifying the first item of interest based on the one or more confidence scores.

12. The system of claim 10 , wherein the activity includes a surge in listings of the first item of interest for sale in the marketplace system.

13. The system of claim 10 , wherein the activity is identified by the machine learning model.

14. The system of claim 10 , wherein the determining the item affinity pattern further comprises:

determining a third set of items that are common to both of the first set of items purchased by the first set of user accounts and the second set of items listed by the second set of user accounts;

determining a second item of interest from the third set of items; and

determining the third set of user accounts associated with listing the second item of interest in the marketplace system.

15. The system of claim 14 , wherein the first set of items excludes the first item of interest, and the second set of items excludes the first item of interest.

16. The system of claim 14 , wherein the determining the item affinity pattern further comprises:

determining a first ranking for the first set of items;

determining a second ranking for the second set of items; and

determining the second item of interest in the third set of items based on the first ranking and the second ranking.

17. The system of claim 16 , wherein the determining the second item of interest in the third set of items based on the first ranking and the second ranking further comprises:

determining, in the third set of items, a particular item having a highest ranking based on the first ranking and the second ranking.

18. The system of claim 10 , wherein the operations further comprise:

causing the notification to be sent to the first set of user accounts associated with the first item of interest.

19. The system of claim 10 , wherein the operations further comprise:

determining a subset of user accounts of the second set of user accounts, the subset of user accounts not being associated with a current listing of the first item of interest for sale; and

causing the notification to be sent to the subset of user accounts of the second set of user accounts.

20. A non-transitory computer-storage medium comprising instructions that, when executed by at least one processor of a machine, cause the machine to perform operations comprising:

using a machine learning model to identify a first item of interest based on an activity in a marketplace system, the machine learning model corresponding to one or more machine learning layers;

determining a first set of user accounts associated with the first item of interest;

determining a second set of user accounts associated with listing the first item of interest for sale;

using the one or more machine learning layers to determine an item affinity pattern between the first set of user accounts and the second set of user accounts, the determining the item affinity pattern being based on a first set of items associated with the first set of user accounts and a second set of items listed by the second set of user accounts;

determining a third set of user accounts based on the first set of user accounts, the second set of user accounts, and the item affinity pattern; and

causing display of a notification on one or more client devices associated with the third set of user accounts, the notification indicating an opportunity to list the first item of interest for sale.