Abstract: A method for targeted advertisement is provided, for which one or more tags relating to an advertisement is/are determined, one or more of the most representative entities for the tag(s) is/are determined, and the advertisement is targeted to the one or more most representative entities. In addition, for each of a plurality of tags, one or more of the most representative entities is/are determined based on term frequency-inverse document frequency, such that an entity is relatively more representative of a tag if the tag is more uniquely and/or frequently associated with the entity. For each tag, the associated entities may be divided into multiple categories, such that one or more most representative entities within each category is/are determined for each tag.

Abstract: A method for targeted advertisement is provided, for which one or more tags relating to an advertisement is/are accessed, one or more of the most representative activities for the tag(s) is/are determined, and the advertisement is targeted to the one or more most representative activities. In addition, for each of a plurality of tags, one or more of the most representative activities is/are determined based on term frequency-inverse document frequency, such that an activity is relatively more representative of a tag if the tag is more uniquely and/or frequently associated with the activity. The advertisement is delivered to participants of the activities, and optionally during a time interval relevant to the activities.

Abstract: A user interface for targeted advertisement is provided, which comprises a plurality of interface components, wherein selected ones of the plurality of interface components are capable of: receiving a list of advertising targets; presenting a list of most representative entities for the list of advertising targets, wherein an entity is relatively more representative of an advertising target if the advertising target is relatively more uniquely and frequently associated with the entity; receiving a list of target entities selected from the list of most representative entities; and presenting an estimated advertising cost for the list of target entities.

Abstract: A method for ranking entities with respect to a tag for targeted advertisement is provided. For each entity to be ranked, a first score is calculated, which indicates the level of uniqueness the tag is to the entity. For each entity to be ranked, a second score is calculated, which indicates the number of people connected with the entity. For each entity to be ranked, a final score is calculated, which combines the first score and the second score. The entities are ranked based on their respective final scores. Optionally, the entity rankings may be used to determine the cost of advertisement for each entity and to recommend entities with respect to an advertisement.

Abstract: Methods, apparatuses and systems directed to creating contextually-targeted advertisements. In a particular implementation, advertisers may leverage a W4 COMN to deliver contextually-targeted and/or contextually-enhanced advertisements. In the implementations discussed below, an ad creation system utilizes data made available by the W4 COMN to facilitate the creation and placement of advertisements on a message delivery network, such as the W4 COMN itself.

Abstract: Methods and apparatus are described by which one or more input words may be predicted based on partial input from a user using a predictive model that employs contextual metadata which characterizes the user in a multi-dimensional space in which the dimensions are defined by one or more of a spatial aspect, a temporal aspect, a social aspect, or a topical aspect.