Synthetic identity detection

Abstract

A system and method for detecting synthetic identities are provided that determine a synthetic identity score for a given user, the synthetic identity score indicating a likelihood that the given user is using a synthetic identity to conduct activities. The synthetic identity score generated by the system and method disclosed herein can then be used to determine a risk associated with the given user and to inform what actions to take based on the associated risk that the given user may use the synthetic identity to perform a bad act.

What is claimed is: 
     
         1 . A system for detecting a synthetic identity, comprising:
 a processor configured to:
 generate a graph network of linked users configured to capture a group of similar applicants in response to user data associated with a given user; 
 derive collective connectivity indicators from the graph network; and 
 determine a synthetic identity score for the given user based at least in part on the collective connectivity indicators, wherein the determined synthetic identity score indicates a likelihood that the given user is using a synthetic identity to conduct activities; and 
   a memory coupled to the processor and configured to provide the processor with instructions.   
     
     
         2 . The system of  claim 1 , wherein the processor is further configured to:
 generate the graph network of linked users by applying a fuzzy clustering technique to soft-link users and a hard clustering technique to hard-link users.   
     
     
         3 . The system of  claim 1 , wherein the processor is further configured to:
 apply rules to the collective connectivity indicators; and   determine the synthetic identity score for the given user in response to the applied rules.   
     
     
         4 . The system of  claim 1 , wherein generating the graph network of linked users includes:
 soft-linking users to the given user in response to a given soft user attribute to form a layer of associated soft-linked users; and   hard-linking users to each soft-linked user in the layer of associated soft-linked users to form a secondary layer of associated hard-linked users.   
     
     
         5 . The system of  claim 4 , wherein:
 each soft-linked user in the layer of associated soft-linked users has one degree of separation from the given user;   each hard-linked user in the secondary layer of associated hard-linked users has two degrees of separation from the given user; and   generating the graph network of linked users includes linking users to the given user in response to a given user attribute to form a layer of associated linked users, wherein each linked user in the layer of associated linked users has one degree of separation from the given user.   
     
     
         6 . The system of  claim 5 , wherein generating the graph network of linked users by:
 hard-linking users to the given user in response to a given hard user attribute to form a layer of associated hard-linked users, wherein each hard-linked user in the layer of associated hard-linked users has one degree of separation from the given user; and   for each linked user in a given layer having two or more degrees of separation from the given user, linking additional users to each linked user in the given layer to form a subsequent layer of subsequently associated linked users to the linked users in the given layer.   
     
     
         7 . The system of  claim 6 , wherein the graph network comprises a plurality of subsequent layers generated by recursively hard-linking additional users to linked users in the graph network in response to a set of shared hard attributes. 
     
     
         8 . The system of  claim 1 , wherein generating the graph network of linked users includes:
 generating a similarity score for a soft-linked user candidate as compared with the given user, wherein a layer of associated soft-linked users is formed by selecting soft-linked user candidates to be soft-linked users in response to similarity scores of the soft-linked user candidates;   hard-linking users to the given user in response to a given hard user attribute to form a layer of associated hard-linked users, wherein each hard-linked user in the layer of associated hard-linked users has one degree of separation from the given user;   hard-linking users to each soft-linked user in the layer of associated soft-linked users to form a secondary layer of associated hard-linked users, wherein each hard-linked user in the secondary layer of associated hard-linked users has two degrees of separation from the given user; and   for each linked user in a given layer having two or more degrees of separation from the given user, hard-linking additional users to each linked user in the given layer to form a subsequent layer of subsequently associated hard-linked users to the linked users in the given layer, wherein the graph network comprises a plurality of subsequent layers generated by recursively hard-linking additional users to linked users in the graph network in response to a set of shared hard attributes.   
     
     
         9 . The system of  claim 8 , wherein the processor is further configured to determine a set of soft-linked users to be soft-linked to the given user by setting a threshold score value and selecting soft-linked user candidates having a similarity score above the threshold score value to be soft-linked users. 
     
     
         10 . The system of  claim 8 , wherein the processor is further configured to determine a set of soft-linked users to be soft-linked to the given user by selecting a specific number of soft-linked user candidates having a highest value of similarity scores to be soft-linked users. 
     
     
         11 . A method of detecting a synthetic identity, comprising:
 generating a graph network of linked users configured to capture a group of similar applicants in response to user data associated with a given user;   deriving collective connectivity indicators from the graph network; and   determining a synthetic identity score for the given user based at least in part on the collective connectivity indicators, wherein the determined synthetic identity score indicates a likelihood that the given user is using a synthetic identity to conduct activities.   
     
     
         12 . The method of  claim 11 , further comprising:
 generating the graph network of linked users by applying a fuzzy clustering technique to soft-link users and a hard clustering technique to hard-link users.   
     
     
         13 . The method of  claim 11 , further comprising:
 applying rules to the collective connectivity indicators; and   determining the synthetic identity score for the given user in response to the applied rules.   
     
     
         14 . The method of  claim 11 , wherein generating the graph network of linked users includes:
 soft-linking users to the given user in response to a given soft user attribute to form a layer of associated soft-linked users; and   hard-linking users to each soft-linked user in the layer of associated soft-linked users to form a secondary layer of associated hard-linked users.   
     
     
         15 . The method of  claim 14 , wherein:
 each soft-linked user in the layer of associated soft-linked users has one degree of separation from the given user;   each hard-linked user in the secondary layer of associated hard-linked users has two degrees of separation from the given user; and   generating the graph network of linked users includes linking users to the given user in response to a given user attribute to form a layer of associated linked users, wherein each linked user in the layer of associated linked users has one degree of separation from the given user.   
     
     
         16 . The method of  claim 15 , wherein generating the graph network of linked users by:
 hard-linking users to the given user in response to a given hard user attribute to form a layer of associated hard-linked users, wherein each hard-linked user in the layer of associated hard-linked users has one degree of separation from the given user; and   for each linked user in a given layer having two or more degrees of separation from the given user, linking additional users to each linked user in the given layer to form a subsequent layer of subsequently associated linked users to the linked users in the given layer.   
     
     
         17 . The method of  claim 16 , wherein the graph network comprises a plurality of subsequent layers generated by recursively hard-linking additional users to linked users in the graph network in response to a set of shared hard attributes. 
     
     
         18 . The method of  claim 11 , wherein generating the graph network of linked users includes:
 generating a similarity score for a soft-linked user candidate as compared with the given user, wherein a layer of associated soft-linked users is formed by selecting soft-linked user candidates to be soft-linked users in response to similarity scores of the soft-linked user candidates;   hard-linking users to the given user in response to a given hard user attribute to form a layer of associated hard-linked users, wherein each hard-linked user in the layer of associated hard-linked users has one degree of separation from the given user;   hard-linking users to each soft-linked user in the layer of associated soft-linked users to form a secondary layer of associated hard-linked users, wherein each hard-linked user in the secondary layer of associated hard-linked users has two degrees of separation from the given user; and   for each linked user in a given layer having two or more degrees of separation from the given user, hard-linking additional users to each linked user in the given layer to form a subsequent layer of subsequently associated hard-linked users to the linked users in the given layer, wherein the graph network comprises a plurality of subsequent layers generated by recursively hard-linking additional users to linked users in the graph network in response to a set of shared hard attributes.   
     
     
         19 . The method of  claim 18 , further comprising determining a set of soft-linked users to be soft-linked to the given user by setting a threshold score value and selecting soft-linked user candidates having a similarity score above the threshold score value to be soft-linked users. 
     
     
         20 . The method of  claim 11 , further comprising determining a set of soft-linked users to be soft-linked to the given user by selecting a specific number of soft-linked user candidates having a highest value of similarity scores to be soft-linked users.