Decision Support Systems

Entity Resolution & Identity Resolution

Entity/Identity resolution defines the known information about an entity (e.g. a person or a location or an object/item) and compiles/associates all known information about that entity until there is enough information to define for certain that it is indeed one entity to which the information relates.

This allows subsequent analysis of all the information to define what is factual information and what is erroneous or false information.

There are significant complexities to this capability, for example people’s names are legitimately written in many ways, especially in Asia and the Middle East.  Even in England and the US name variations can be valid e.g. Richard, Rich, Dickie.  Add to this the variety of data storage methods and the number of legitimate variant names for each person increases substantially.  The use of pre-fixes such as Mr, Mrs, Ms, Miss, Dr. Rev. and the ability to use initials, full names, middle names and maiden names adds to the number of apparent ‘identities’ that can relate to the same individual.

Entity/identity resolution analyses all this information to ascertain links and associations which are likely to indicate that differing information/data is likely to refer to the same person or entity.  This could be from other factors – in the case of Identity Resolution for a person, their address, date of birth, social security number or other key data.  This is not always conclusive as, for example, father and son could have the same initial and even first and family names and live at the same address – therefore further data would be require to identify one from the other.

Add in the complexity of individuals who deliberately falsify information for criminal purposes and the problem becomes more challenging.

It is worth noting that this is not the same as master data management (MDM).  All the information is valid and so although one aims to provide a ‘single version of the truth’ the ‘truth may well include multiple values for the same attribute – especially names and addresses.  Therefore the objective is not to normalize or standardize formats but to be certain when different versions of a legitimate attribute relate to the same individual or entity and provide the consolidated view of that entity with all the various values for each attribute registered or linked to the individual or entity.

It is also important to be able to link together differing attributes which suggest different individuals but which are actually the same individual i.e. when dates of birth are written US format or UK format or when the individual is deliberately attempting to conceal their true identity.  Thus one generates a view of all the versions of attributes which an individual uses to disguise their true identity and compile these into one entity record so that the complete picture of the individual of interest can be collated as if it were one ‘record’.  Combine this with Relationship Analysis to understand the links between individuals/entities and other entities.

For further information on decision support systems contact us.