System and method for web destination profiling
An improved system and method for web destination profiling for online population-targeted advertising is provided. A web destination profiler may be provided for generating web destination profiles. Traffic may be analyzed at a particular web destination in order to understand the population visiting the web destination. The analysis of user traffic, including differentiated clickstream data, may be applied for determining known characteristics of a web destination profile. Moreover, unknown characteristics of a web destination profile may be determined using a variety of techniques including inferring characteristics by modeling traffic flow through other web destinations, estimating characteristics from other web destination profiles by predicting traffic flow through other web destinations, propagating characteristics to a web destination profile by smoothing a joint distribution of characteristics of other web destination profiles, and so forth. Web destination profiles may be used by applications such as an online application for population-targeted advertising.
1. A computer-implemented method for profiling a web destination, comprising:
using a processor device, performing:
analyzing traffic at a web destination for determining unknown characteristics of a population of visitors to said web destination, said traffic comprising both differentiated and undifferentiated clickstream trails, wherein said undifferentiated clickstream trails comprise clickstream data that is undifferentiated as to the visitors to the web destination;
generating a distribution of characteristics of the population of visitors to the web destination;
generating a web destination profile over the distribution of characteristics;
propagating the web destination profile with the distribution of characteristics from profiles of other web destinations by following the clickstream trails; and
storing the web destination profile in storage.
2. The method of claim 1 further comprising determining known characteristics of the population of visitors to the web destination for the web destination profile.
3. The method of claim 1 wherein determining the unknown characteristics of the population of visitors to the web destination for the web destination profile comprises inferring at least one characteristic to the web destination profile from other web destination profiles by modeling traffic flow through other web destinations.
4. The method of claim 1 wherein determining the unknown characteristics of the population of visitors to the web destination for the web destination profile comprises estimating at least one characteristic of the web destination profile from other web destination profiles by predicting traffic flow through other web destinations.
5. The method of claim 4 wherein estimating the at least one characteristic of the web destination profile from other web destination profiles by predicting traffic flow through other web destinations comprises determining a sum of a distance between a predicted distribution and an observed distribution and a distance between the predicted distribution and a propagated distribution determined by the traffic arriving at the web destination.
6. The method of claim 4 wherein estimating the at least one characteristic of the web destination profile from other web destination profiles by predicting traffic flow through other web destinations comprises determining a maximal entropy for an entropy model of the other web destinations.
7. The method of claim 1 wherein determining the unknown characteristics of the population of visitors to the web destination for the web destination profile comprises propagating at least one characteristic to the web destination profile by smoothing a joint distribution of characteristics of other web destination profiles.
8. The method of claim 7 wherein propagating the at least one characteristic to the web destination profile by smoothing the joint distribution of characteristics of other web destination profiles comprises smoothing a joint distribution of characteristics of other web destination profiles for web destinations visited by a user in a session.
9. The method of claim 7 wherein propagating the at least one characteristic to the web destination profile by smoothing the joint distribution of characteristics of other web destination profiles comprises smoothing a joint distribution of characteristics of other web destination profiles for web destinations selected by topical similarity of content.
10. The method of claim 7 wherein propagating the at least one characteristic to the web destination profile by smoothing the joint distribution of characteristics of other web destination profiles comprises smoothing a joint distribution of characteristics of other web destination profiles for web destinations selected by similarity of clickthrough behavior of visitors.
11. A non-transitory computer-readable medium having computer-executable instructions for performing the method of claim 1 .
12. A computer system for profiling a web destination, comprising:
storage storing a web destination profile;
memory with computer-executable instructions stored thereon; and
a processor device operably coupled with the memory, executing the computer-executable instructions comprising:
analyzing traffic at the web destination for determining at least one unknown characteristic of a population of visitors to said web destination, the traffic comprising both differentiated and undifferentiated clickstream trails, wherein said undifferentiated clickstream trails comprises clickstream data that is undifferentiated as to the visitors to the web destination;
determining a distribution of characteristics of the population of visitors to the web destination;
generating the web destination profile over the distribution of characteristics for the population of visitors; and
propagating the web destination profile with the distribution of the characteristics from profiles of other web destinations by following the clickstream trails.
13. The system of claim 12 wherein the computer-executable instructions further comprise:
analyzing the clickstream data from the web destination.
14. The system of claim 12 further comprising a topology analysis engine for analyzing topology data about the web destination.
15. The system of claim 12 further comprising a model generator for generating a model of traffic flow among a plurality of web destinations.
16. The system of claim 15 further comprising a traffic flow analysis engine for propagating population characteristics to a plurality of web destination profiles by predicting traffic flow through the plurality of web destinations.