Abstract: Methods and apparatus for managing models for classification of online users are disclosed. An example method includes determining a segment associated with a received impression, determining whether the received impression includes an indication that the received impression is the result of a conversion associated with the segment, determining a current ratio of converted to non-converted impressions associated with a classification model associated with the segment, determining a difference between the current ratio and a target ratio, and storing an indication of the received impression for use in at least one of training or updating a classification model associated with the segment when: (a) the difference does not meet a threshold or (b) counting the received impression in the ratio would reduce the difference.

Abstract: A targeting rule set of association rules may be created. A plurality of association rules may be selected. A plurality of candidate rule sets may be constructed based on the plurality of association rules. The plurality of candidate rule sets may be evaluated to produce metrics. One of the plurality of candidate rule sets may be assigned as the targeting rule set based on the metrics. A custom list of user may be formed using association rules. A user may be determined to belong to a segment by applying one or more rules of a targeting rule set to user attribute data. A custom list of cookies to show advertising may be formed using combinations of association rules.

Abstract: Methods, system, and media for scaling a panel of users with known attributes to determine an unknown attribute of a user are disclosed. An unclassified attribute vector of the user may be received. The unclassified attribute vector may comprise first attributes and first attribute values. A reduced attribute vector may be produced using the unclassified attribute vector and a projection matrix. The reduced attribute vector may have fewer attributes than the unclassified attribute vector. A plurality of reduced panel attribute vectors may be produced using a plurality of classified panel attribute vectors of users from the panel and the projection matrix. The reduced panel attribute vectors may have fewer attributes than the classified panel attribute vectors. The unknown attribute of the user may be determined based on the reduced attribute vector, the plurality of reduced panel attribute vectors, and the known attributes.

Abstract: Methods, systems, and media for determining an index of a segment are disclosed. A first value representing one or more first users in a segment that have performed one or more key events may be determined. A second value representing one or more second users of a population that have performed the one or more key events may be determined. A third value representing one or more third users in the segment that have not performed the one or more key events may be determined. A fourth value representing one or more fourth users of the population that have not performed the one or more key events may be determined. The index for the segment may be calculated based on the first value, the second value, the third value, and the fourth value. The index may be transmitted.

Abstract: In a network comprising number of web sites and at least one simulator web page calls are simulated or monitored and the response to the calls, including content, re-directed calls and cookies are examined in order to identify whether the content of cookies has been written or modified illegally. Illegal modification of content of cookies is referred to as writing of data by a Buyer in cookies of a User that was directed to the Buyer by a Data Publisher, without the consent of the Data Publisher. In some embodiments when illegal cookies are identified an alert may be issued to a user.

Abstract: Methods, system, and media for scaling a panel of users with known attributes to determine an unknown attribute of a user are disclosed. An unclassified attribute vector of the user may be received. The unclassified attribute vector may comprise first attributes and first attribute values. A reduced attribute vector may be produced using the unclassified attribute vector and a projection matrix. The reduced attribute vector may have fewer attributes than the unclassified attribute vector. A plurality of reduced panel attribute vectors may be produced using a plurality of classified panel attribute vectors of users from the panel and the projection matrix. The reduced panel attribute vectors may have fewer attributes than the classified panel attribute vectors. The unknown attribute of the user may be determined based on the reduced attribute vector, the plurality of reduced panel attribute vectors, and the known attributes.

Abstract: System and method for automatic definition and tuning of user segment combinations used for targeted advertisement and marketing. Automated analysis and feedback modules may provide, ad hoc, optimized market segments definitions that maximize the conversion ratio for advertisements. The system and method may optimize of segment data based on a learning-running process built around a data exchange within a real-time bidding (RTB) media exchange system. In identifying qualified combinations for segments on-going loops or iterations may analyze the correlation between the performance of an advertisement or “creative” and user data, based on a set of parameters. As a part of data segmentation, embodiments of the present invention find combinations of parameters that may predict lift. Segment combinations are optimized by, for example combining a segment combination with one or more other segment combinations and/or dividing a segment combination into two or more different combinations.

Abstract: A targeting rule set of association rules may be created. A plurality of association rules may be selected. A plurality of candidate rule sets may be constructed based on the plurality of association rules. The plurality of candidate rule sets may be evaluated to produce metrics. One of the plurality of candidate rule sets may be assigned as the targeting rule set based on the metrics. A custom list of user may be formed using association rules. A user may be determined to belong to a segment by applying one or more rules of a targeting rule set to user attribute data. A custom list of cookies to show advertising may be formed using combinations of association rules.

Abstract: A method, system, and medium for managing modeled audience extension information that includes receiving tag information, the tag information comprising a user identifier corresponding to a user, a tag identifier, and one or more data values; adding the tag identifier and the one or more data values to a user object, wherein the user object is identified by the user identifier; assigning the user object to a segment based on the user object and a model, wherein the segment is identified by a segment identifier; and distributing the segment identifier and the user identifier to a market information buyer.

Abstract: A method is disclosed which emulates accessing web sites by a simulated web visitor. The characteristics of the web user may be controlled as may be the timing and number of visits and the nature of the web sites visited. During the emulated visits cookies of the web visitor are modified. The emulated visits may be directed to a sub-group of web sites which are engaged in agreements granting them reward for the provision of data indicative of the nature of past web visits, and to a sub-group of web sites which are not engaged in such agreements. Comparison of the exposure of the emulated web visitor to advertisements in each of the sub groups may be analyzed to indicate whether the agreements are properly executed and the reward is properly given.

Abstract: System and method for automatic definition and tuning of user segment combinations used for targeted advertisement and marketing. Automated analysis and feedback modules may provide, ad hoc, optimized market segments definitions that maximize the conversion ratio for advertisements. The system and method may optimize of segment data based on a learning-running process built around a data exchange within a real-time bidding (RTB) media exchange system. In identifying qualified combinations for segments on-going loops or iterations may analyze the correlation between the performance of an advertisement or “creative” and user data, based on a set of parameters. As a part of data segmentation, embodiments of the present invention find combinations of parameters that may predict lift. Segment combinations are optimized by, for example combining a segment combination with one or more other segment combinations and/or dividing a segment combination into two or more different combinations.

Abstract: A system and method for providing a targeting event marketplace is provided. Generally, the method contains the steps of: at least one targeted publisher and a targeting event marketplace entity agreeing to financial terms associated with revenues received from at least one buyer that obtains access to the Internet browser of an end-user; providing an end-user tag on a Web page of the Web site, wherein the end-user tag is capable of calling a Web server from an Internet browser of the end-user; analyzing an end-user action associated with the Web page, wherein the step of analyzing is performed to categorize the action into a category of targeting event; the Web server determining if at least one buyer has interest in at least one end-user taking an action that is categorized into at least one category of targeting event; and receiving bids from at least one buyer for providing access to the Internet browser of the end-user and selecting at least one buyer.