×

Application of user context to searches in a virtual universe

  • US 8,271,475 B2
  • Filed: 05/27/2008
  • Issued: 09/18/2012
  • Est. Priority Date: 05/27/2008
  • Status: Expired due to Fees
First Claim
Patent Images

1. A method, performed on a computer system, for performing a search for a resource in a virtual universe using user context, the method comprising:

  • using the computer system to perform the following;

    receiving a query from an avatar that is online in the virtual universe;

    scanning a collection of avatar data describing attributes that are relevant to behavioral, search and informational needs of the avatar, wherein the scanning of a collection of avatar data comprises scanning all of the following;

    inventory items belonging to the avatar, teleportation history of the avatar, motion history of the avatar and social tagging behavior exhibited by the user of the avatar in the real world;

    using the scanned collection of avatar data to determine a user context for the avatar from at least one of past behavior in the virtual universe or past behavior exhibited by the user of the avatar in the real world, wherein the determining of a user context for the avatar comprises using a plurality of matching techniques to assign a mutually exclusive category designation from a list of mutually exclusive category designations applicable to each scanned collection of avatar data including the inventory items belonging to the avatar, the teleportation history of the avatar, the motion history of the avatar and the social tagging behavior exhibited by the user of the avatar in the real world, to specific avatar data obtained therefrom and comparing the specific avatar data with criteria associated with the assigned category designation to identify a value with an attribute that provides user context of the avatar, wherein the determining of a user context for the avatar further comprises using machine learning techniques, wherein the machine learning techniques comprise an unsupervised machine learning technique that discovers and updates user contexts from the past behavior in the virtual universe and the real world and a supervised machine learning technique that refines user context attributes and values based on user interactions with a virtual universe search tool, wherein the unsupervised machine learning technique clusters avatar data according to a distance metric to determine user contexts, associated attributes and values for the attributes and updates the user contexts, associated attributes and values as more avatar data becomes available, wherein the unsupervised machine learning technique leverages user interaction data of all avatars in the virtual universe by enriching data associated with the avatar performing the search with data associated with similar avatars as determined by clustering avatar data; and

    performing a resource search for the query in accordance with one of the user contexts determined from the scanned collection of avatar data.

View all claims
  • 1 Assignment
Timeline View
Assignment View
    ×
    ×