Abstract:
An online system receives information from an entity identifying a set of users of the online system and groups users included in the set into clusters based on their similarities using a clustering model or algorithm (e.g., k-means clustering) and based on one or more parameters specified by the entity. The online system generates expanded clusters that include additional users in one or more clusters based on similarities between the additional users and users in various clusters. If an additional user is included in multiple expanded clusters, the online assigns the additional user exclusively to an expanded cluster that best fits the user.
Abstract:
A target audience for an ad campaign is determined during an exploration period of the ad campaign by modifying the target audience based on the fulfillment of performance objectives. An initial target audience may be provided by the advertiser or determined by the social networking system based on ad campaigns having similar ad content or other similar characteristics. Advertisements associated with the ad campaign are served to users of the initial target audience. A subset of the target audience that fulfills the performance objectives of the ad campaign is identified and those users are used to generate a new targeting audience to target users that “look like” the subset of the target audience. The new targeting audience is used in place of the initial target audience to improve targeting for the advertisement. This process may be iteratively performed to refine the target audience during the exploration period.