METHODS AND SYSTEMS FOR MINING WEBSITES

• US 20100121850A1
• Filed: 11/10/2008
• Published: 05/13/2010
• Est. Priority Date: 11/10/2008
• Status: Active Grant

• Current Assignee: NBC Universal Media LLC (Comcast Corporation)
• Original Assignee: General Electric Company
• Inventors: Gustafson, Steven Matt, Moitra, Abha, Xue, Feng
• Granted Patent: US 8,219,583 B2
• Time in Patent Office: Days
• US Class Current: 707/737
• CPC Class Codes: G06F 16/95 Retrieval from the web
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

Mining of websites that in one embodiment includes obtaining web usage data of user sessions of a website, wherein the website has a hierarchical structure with granular levels and has mapping from each webpage of the website into the hierarchical structure, mapping the user sessions to the hierarchical structure of the website resulting in hierarchical user sessions, initiating an edit distance metrics to determine similarity in the hierarchical user sessions, and clustering similar hierarchical user sessions into groups.

Claims

1. A method for mining websites, comprising:

obtaining web usage data of user sessions of a website, wherein said website has a hierarchical structure with granular levels and has mapping from each webpage of the website to the hierarchical structure;

mapping the user sessions to the hierarchical structure of the website resulting in hierarchical user sessions;

initiating metrics to determine similarity in the hierarchical user sessions; and

clustering similar hierarchical user sessions into groups.

2. The method of claim 1 wherein the hierarchical structure includes differentiating content of a website in at least two levels of hierarchy.

3. The method of claim 1 wherein the web usage data is obtained for a fixed period of time.

4. The method of claim 3 wherein the fixed time period is customized as per business interest.

5. The method of claim 1 wherein said user session is for a time period from the start time of a user activity on the website till time of exit of user from the website.

6. The method of claim 1 wherein the user session is a sequence of page views.

7. The method of claim 1 wherein the user session includes at least one of paths of page views, sessions of same users, or sessions of different users.

8. The method of claim 1 wherein said metrics determines a cost of conversion of one hierarchical session to another hierarchical session.

9. The method of claim 1 wherein the metrics determines a minimum number of operations performed to transform web surfing paths of a hierarchical user session into web surfing paths of another hierarchical user session.

10. The method of claim 1 wherein the metrics determine at least one of a distance between paths of two or more hierarchical user sessions or a distance between each pair of hierarchical user sessions.

11. The method of claim 10 wherein the distance between paths of two different hierarchical user sessions is determined using one or more of Levenshtein Distance, Damerau-Levenshtein distance, Jaro-Winkler distance, Jaro and McBrideQuimbyShih distance.

12. The method of claim 1 further comprising visualizing clusters and detecting trends or anomalies.

13. A method for segmenting website users, comprising:

obtaining web usage data of said users of a website, wherein said website has a hierarchical structure with granular levels and has mapping from each webpage of the website to the hierarchical structure;

characterizing said users to determine user vectors of the users;

clustering said users into groups based upon similarity in user interest;

determining centroids of the groups using the user vectors of each of the groups wherein the centroid represents interest vectors of the groups; and

determining closeness between the groups or atleast one user and atleast one group.

14. The method of claim 13 wherein characterizing user sessions comprises:

defining user vectors for each user including number of page views in each level of hierarchy;

incorporating business preferences in the user vectors; and

determining normalized vectors by dividing page views in each level by total number of page views by the user.

15. The method of claim 13 wherein the hierarchical structuring includes differentiating the content of a website into at least two levels of hierarchy.

16. The method of claim 13 wherein the levels include categories or subcategories of content of the website.

17. The method of claim 13 wherein page views of the website users are collected for each user.

18. The method of claim 13 wherein page views of the website are collected for selected users.

19. The method of claim 13 wherein user vectors comprise a number of page views by the user in each level of hierarchy.

20. The method of claim 14 wherein the normalized vector is determined by dividing number of page views of each level of a user by a total number of page views by the user.

21. The method of claim 13 wherein the closeness is performed by determining similarity of normalized vector of one said user and ranking the closeness of the user to the existing groups.

22. The method of claim 13 wherein the distance between centriods is determined using a Mahalanobis distance method.

23. The method of claim 14 wherein the number of page views is altered in each user vector depending on a business preferences.

24. The method of claim 13 wherein the page views of users are collected for a configurable predefined period of time.