NAS (Network Analysis Server)

 

NAS is the SN-Sphere application server which provides analytical services Reveal , Consider, Contextor  and all other components of SN-Sphere. NAS services are also available to third party clients using web services.

In addition to Network Pattern Analysis and Social Network Analysis executed by NAS and implemented in  Reveal, NAS can currently execute several types of analysis: network link analysis, connectivity analysis and similarity analysis . Additional types of analysis are under development.

Network Link Analysis

Network link analysis uses LSI (Link Support Information) to infer links between entities in order to expose and analyze network associations. A network is composed of one or more types of entities such as people, organizations and financial organizations. Different LSI may be used for inferring links between the different combinations of entities. The following matrix is a simple illustration of this concept for an anti-money laundering (AML) application. The same principles can be applied for any application using various types of data.

Each cell indicates examples of LSI which may be used for inferring links between specific types of entities. An AML application may analyze many types of entities using many sources of LSI.


 

Financial Institution

Organization

Person

 

Account owner/client

Withdrawal

Deposit

Transfer

Employed by

Pension Fund

Mortgage

Ownership

Employed by

Telephone number

Address

SAR (Suspicious Activity Report)information

Joint bank account

Family related

SAR (Suspicious Activity Report)information

Telephone number

Address

Monetary transfers

Person

Account owner/client

Withdrawal

Deposit

Transfer

Ownership

Loan

Monetary transfer

Business relationship

Ownership

SAR (Suspicious Activity Report)information

 

Organization

Transfer

Ownership

Consortium

 

 

Financial Institution

 

A suspicious individual, company, account or other starting point (target) is chosen for analysis. The user may choose one or more types of LSI (Link Support Information) and define one or more relevant criteria for the analysis. For instance, a criterion for inferring links between two people based on monetary transfers might be that they both transferred money to the same third party within a certain period of time in amounts exceeding a particular threshold.

NAS uses the chosen criteria and LSI to infer links between the target and other entities (first generation of direct associations) as well as cross links between the members of that generation. The system then continues a recursive analysis in which further links are inferred between each of the entities in the first generation and other entities, thus creating a second generation. The process is continued for as many generations as the user has requested. The end result is a network of entities: who is involved and how are they related. The results can be displayed in a link chart and/or grid format.

 


























Until now, the use of link analysis has been limited due to the excessive amount of time it takes to execute such analysis when using even several tens of thousands of data records let alone millions or billions. Svivot has developed proprietary unique technology, called PreStar, which allows the link analysis to be completed in a very short amount of time even when using massive amounts of data. NAS is specially designed to be able to work with PreStar.

Connectivity Analysis

Connectivity analysis determines if a specific entity is related to another single entity or group of entities in some way even if the relationship is indirect through many degrees (generations) of separation. The system can find the shortest path of connectivity between the entities or all paths by which they are connected. NAS uses selected LSI for the analysis. Connectivity analysis can use all of the available link support information, only link that appear on previously saved link charts or only links on a single link chart. The result of the analysis is a new link chart displaying the requested connectivity.

Similarity Analysis

Similarity analysis determines if two specific entities or a list of entities are similar to each other and if so in what ways. Similarity is determined by:

  • Comparison of entity attributes - such as for a person, attributes such as address, phone number, credit card, hair color, date of birth, SSN, etc. The degree of similarity can be calculated according to how closely the attributes match and based on multiple attributes

  • Comparison of associated entities - For instance, credit cards which are suspected of having been used fraudulently could be considered similar if they were used at the same point of sale within a certain period of time. 

  •