Recommending content items

摘要

Recommending content items is disclosed. In some embodiments, determining an item to recommend includes receiving an indication that a first item is of interest and selecting a second item to recommend based at least in part on a degree of similarity of the second item to the first item as determined based at least in part on a number of other instances in which both the first item and the second item have been of interest and a first decayed popularity of the first item. In some embodiments, the degree of similarity of the second item to the first item is determined based in addition on a second decayed popularity of the second item.

主权项

1. A method of determining an item to recommend, comprising:
receiving at a processor an indication that a first item is of interest to a user; determining a second item to recommend to the user based at least in part on a degree of similarity of the second item to the first item; and providing the second item as a recommendation to the user in the event that the second item is not already associated with the user; wherein the degree of similarity is determined based at least in part on a similarity score that quantifies the degree of similarity of the second item to the first item based at least in part on a number of other instances in which both the first item and the second item have been of interest, a decayed popularity of the first item, and a decayed popularity of the second item; wherein when computing a particular instance of the similarity score, for each of the first item and the second item, an instance the item has been of interest to another user is given a value that is decayed over time by a non-zero decay factor whose value is based at least in part on a time at which the instance occurred such that a more recent instance in which the item has been of interest to another user contributes more to the decayed popularity of that item than a less recent instance in which that item has been of interest to another user so that the decayed popularity of the item is biased towards more recent instances of interest in the item by one or more other users.