Incremental update of long-term and short-term user profile scores in a behavioral targeting system
View Patent ↗A behavioral targeting system determines user profiles from online activity. The system includes a plurality of models that define parameters for determining a user profile score. Event information, which comprises on-line activity of the user, is received at an entity. To generate a user profile score, a model is selected. The model comprises recency, intensity and frequency dimension parameters. The behavioral targeting system generates a user profile score for a target objective, such as brand advertising or direct response advertising. The parameters from the model are applied to generate the user profile score in a category. The behavioral targeting system has application for use in ad serving to on-line users.
1. A method for determining a user profile from online activity, said method comprising:
collecting event information up to a first period of time, wherein said event information comprises on-line activity between a user and an entity;
processing said event information so as to generate a long-term user profile score including at least one of computing an intensity dimension comprising a decay component or computing a recency dimension, wherein said long-term user profile score is used to assess user behavior based on said user's long-term interest;
receiving, at said entity, event information subsequent to said first period of time;
processing said event information received subsequent to said first period of time so as to generate a short-term user profile score, wherein said short-term profile score is used to assess user behavior based on said user's short-term interest;
re-calculating said long-term user profile score without re-processing said event information collected up to said first period of time by either increasing said decay component based on a time associated with said user profile score or incrementing said recency dimension based on a time associated with said user profile score; and
combining said long-term user profile score and said short-term user profile score to generate a user profile score.
2. The method as set forth in claim 1 , wherein processing said event information so as to generate a long-term user profile score comprises batch processing said event information to generate said long-term user profile score.
3. The method as set forth in claim 1 , wherein processing said event information to generate a short-term user profile score comprises processing said event information in real-time to generate said short-term user profile score.
4. The method as set forth in claim 1 , wherein said time subsequent to said first period of time comprises at least one hour.
5. The method as set forth in claim 1 , wherein said time subsequent to said first period of time comprises two days.
6. A computer readable storage medium comprising a set of instructions which, when executed by a computer, cause the computer to determine a user profile from online activity, said instructions for:
collecting event information up to a first period of time, wherein said event information comprises on-line activity between a user and an entity;
processing said event information so as to generate a long-term user profile score including at least one of computing an intensity dimension comprising a decay component or computing a recency dimension, wherein said long-term user profile score is used to assess user behavior based on said user's long-term interest;
receiving, at said entity, event information subsequent to said first period of time;
processing said event information received subsequent to said first period of time so as to generate a short-term user profile score, wherein said short-term profile score is used to assess user behavior based on said user's short-term interest;
re-calculating said long-term user profile score without re-processing said event information collected up to said first period of time by either increasing said decay component based on a time associated with said user profile score or incrementing said recency dimension based on a time associated with said user profile score; and
combining said long-term user profile score and said short-term user profile score to generate a user profile score.
7. The computer readable storage medium as set forth in claim 6 , wherein processing said event information so as to generate a long-term user profile score comprises batch processing said event information to generate said long-term user profile score.
8. The computer readable storage medium as set forth in claim 6 , wherein processing said event information to generate a short-term user profile score comprises processing said event information in real-time to generate said short-term user profile score.
9. The computer readable storage medium as set forth in claim 6 , wherein said time subsequent to said first period of time comprises at least one hour.
10. The computer readable storage medium as set forth in claim 6 , wherein said time subsequent to said first period of time comprises two days.
11. A system comprising:
at least one server for collecting a first set of event information up to a first period of time, and for receiving a second set of event information subsequent to said first period of time, wherein said event information comprises on-line activity between a user and an entity;
batch processing unit, coupled to said server, for processing said event information so as to generate a long-term user profile score including at least one of computing an intensity dimension comprising a decay component or computing a recency dimension, wherein said long-term user profile score is used to assess user behavior based on said user's long-term interest;
real-time processing unit, coupled to said server, for processing said event information received subsequent to said first period of time so as to generate a short-term user profile score, wherein said short-term profile score is used to assess user behavior based on said user's short-term interest; and
wherein, said server for re-calculating said long-term user profile score without re-processing said event information collected up to said first period of time by either increasing said decay component based on a time associated with said user profile score or incrementing said recency dimension based on a time associated with said user profile score, and for combining said long-term user profile score and said short-term user profile score to generate a user profile score.
12. The system as set forth in claim 11 , wherein said time subsequent to said first period of time comprises at least one hour.
13. The system as set forth in claim 11 , wherein said time subsequent to said first period of time comprises two days.