ELISE ID Platform

Identity Matching

WCC’s identification platform combines biographic and biometric data to offer fast identification capabilities. The ELISE ID Platform searches intelligently, using built-in fuzzy logic algorithms and weighted criteria in both structured and unstructured data. As well as from disparate databases and diverse sources too.

English

Spanish

Portuguese

ELISE ID Platform Search & Match

It’s important to realize that Identity and Security applications require new technologies in order to identify risks accurately. Firstly, because the data used for matching may be incomplete or inaccurate. Secondly, sources may present conflicting versions of the same information. There may be different versions of the same name due to regional differences, linguistics, or typos. Then again, there may be differences in the way data is collected and stored in multiple databases.

In reality, such challenges are present in every security application. Whether for recognizing or registering legal citizens, or of course to identify criminals, criminal activities, terrorists, and illegal immigrants during border controls.

So, building on more than 25 years of experience in the Identity and Security sector, WCC developed the ELISE ID Platform. Our MITRE award-winning and unique Search & Match engine that lies at the center of our Identity and Security solutions: HERMES and HERA.

ELISE ID Platform Search & Match

Identity Management Platform

Out-of-the-box solutions for Border Management and Civil Registration

Challenges to making a good match

Matching data in the real world can often present significant challenges. For one thing, watch lists frequently contain poor quality or incomplete data. Partly because mistakes may be made when entering data. But also because of information gaps where information is unknown to the officer entering the data.

As can be seen, typographical mistakes are possible in all types of attributes: names, dates, city names, or country names. But as a result, good quality matching can be difficult. And significant matches will therefore be lost. Besides typographical errors, other differences can also influence the matching process. For example, name variations (Alexander, Iskander, Oleksandr à Sasha) play a role. As well as multilingual city and country names. In other words, it is essential to use smart software that allows for such variations.

Name
challenges

Typing errors
John > Jhon
Transcription variants
Mohammad > Muhamet
Name variations
William > Bill
Name changes
After marriage, adoption

Date of birth challenges

Typing errors
11-08-1975 vs 11-08-197
DD-MM vs MM-DD
11-08-1975 vs 08-11-1975

City name challenges

Typing errors
Multilingual names
The Hague, Den Haag, 's Gravenhage, La Have

Country name challenges

Typing errors
Multilingual names
The Netherlands, les Pays Bas, Holland, Nederland
Geographic changes
former USSR, Russia, Ukraine, Belarus

ELISE ID Platform – similarity matching

In spite of the above, traditional databases and search platforms don’t offer intelligent search and match capabilities. Instead, they use queries based on precise input. And so return exact output. In other words, straightforward Boolean logic, representing an exact match. But what about a situation when the data entered is inaccurate? Similarly when there are errors in the data stored? Or what if incorrect assumptions are made about how the data is represented. Or even the language and character set used? In these examples, critical search criteria will return only a few results. Or on the other hand, perhaps none at all. While at the same time, less specific criteria will return many irrelevant results.

By contrast, the ELISE ID Platform uses fuzzy logic to provide meaningful match results. Employing multiple, detailed criteria does not, in fact, limit the results. On the contrary, this approach delivers results of higher quality. Because of our support for fuzzy matching, and the use of similarity ranges, ELISE ID Platform will significantly enhance the search quality. And equally important, it will improve the performance too. As an illustration, take a look at the following examples:

Height

Name matching

Similarities

Distance

Date of Birth

Height

Name matching

Similarities

Distance

Date of Birth

In each example, a range can express the degree of similarity between the search criteria and the data. The age of an offender or a victim is often an estimate. So matches are made within a range and scored appropriately. A similar approach is possible for hair color, distance, heights, or names. With this in mind, it’s easy to see how the ELISE ID Platform offers a unique advantage over traditional technology.

ELISE ID Platform – just how does it work?

In the first place, ELISE ID Platform stores all identity-related data within a single object. Then ELISE evaluates each separate search criterion against the property in each data object. As a result, it calculates a match score. ELISE then intelligently combines each score, to calculate the total match score for that identity. And so we can indicate how well it meets the given search criteria. After completing an exhaustive search on all records, ELISE presents the best matches in ranked order. Finally, we can apply a cut-off value. Therefore, we can restrict the number of matches, using either a maximum number of hits (top n) or a minimum specified match score (threshold). What is more, ELISE can also match complex data types, like biometric matching, face recognition, or fingerprint matching. To this end, external algorithms can seamlessly be integrated as ELISE ID Platform plugins.

Key ELISE ID Platform matching features

Different weighting of each criterion

Not all criteria are equally important. So ELISE ID Platform lets users define default weights. Or instead they can influence weights at query time.

Built-in functionality for unstructured text matching

Dynamically controls the importance of each keyword, such as name parts. Also shows users exactly how their query was interpreted.

Fuzzy matching on words, concepts and numeric quantities

Flexible numeric boundaries make it possible to find dates or other quantitative values that are just outside the ideal range. Affinity sets also mean similar or related concepts, words or names are found.

Core identity services like enrollment, identification, and deduplication

Supports various types of biometrics, such as Fingerprint, Iris, Face Recognition, and Voice.

Analytics to understand match scores and result ranking

Transparent calculations behind the match scores in order to add insight into result ranking.

Deterministic and repeatable match scoring

Match score calculations are based on known algorithms, so results can be repeated and refined.

Deterministic match behavior, following the rules laid out in the fully configurable ELISE Dictionary

The Dictionary can quickly be reloaded on-the-fly, so there is no downtime. Furthermore, most of its rules can be overridden as required at query time.

Key ELISE ID matching features:

Different weighting of each criterion
Not all criteria are equally important. So ELISE ID Platform lets users define default weights. Or instead they can influence weights at query time.
Built-in functionality for unstructured text matching
Dynamically controls the importance of each keyword, such as name parts. Also shows users exactly how their query was interpreted.
Fuzzy matching on words, concepts and numeric quantities
Flexible numeric boundaries make it possible to find dates or other quantitative values that are just outside the ideal range. Affinity sets also mean similar or related concepts, words or names are found.
Core identity services like enrollment, identification, and deduplication
Supporting various types of biometrics, such as Fingerprint, Iris, Face Recognition, and Voice.
Analytics to understand match scores and result ranking
Transparent calculations behind the match scores in order to add insight into result ranking.
.

Matching features

In ELISE ID Platform, a query used to find a suspect is in fact a data object itself. These query data objects can be therefore be stored in ELISE ID Platform so they are available for reuse. Users can then revisit previous searches, and search within these stored queries.

Matching Features

ELISE ID Platform can use fuzzy logic to find results that match the query most closely. There are several types, such as numeric ranges, values that have an affinity, text searches that correct for typos, and biometric scores which are never 100%.

Matching Features

in the ELISE ID Platform Dictionary, users can store affinities between words or phrases that are similar in meaning (e.g. “ship”, “vessel”, “boat”, “yacht”). In addition, they can share the same word root (e.g. “clean”, “cleaner”, cleaning”) . WCC also provides the Affinity Generator, which can generate word affinities based on large bodies of text.

Matching Features

ELISE ID Platform matches strings and fully supports Unicode. So any type of textual data, such as a license plate or a passport number, can be entered and searched.

Matching Features

ELISE ID Platform can predict user input based on prior input, a list of values, and/or a taxonomy. Other means of predictive searching are also possible, but will be part of the integration effort.

Matching Features

Users can quickly search for a number within a numerical range. Ranges can be gliding, so for example an age might be given as “40 to 45, maybe 37, but definitely not older than 50”.

Matching Features

ELISE ID Platform supports matches on value ranges. For example, a text attribute representing a serial number. Such “numbers” usually contain letters too (e.g. NS265056), so a simple numerical range search cannot be used for this kind of attribute. With ELISE ID Platform, users can also define value ranges for text properties. Textual data can be stored or requested as a range, for example from 3400AA to 3400ZZ. ELISE ID Platform will be able to identify an item (e.g. 3400KN) in such a range. Similarly, NS265056 will be identified as within the range NS265000 – NS265500.

Matching Features

In text searching, ELISE ID Platform recognizes:

regular expressions like * and ?

extended wildcards like {1,4}

logical and text operators like AND, INT, OR, NOT, + and –

Matching Features

Users can store aliases and nicknames in the ELISE ID Platform Dictionary as text affinities. The ELISE ID Platform Dictionary also fully supports Unicode. So names in any script can be stored and searched using ELISE ID Platform’s transcription capability.

Matching Features

IIf required, ELISE ID Platform can easily integrate third-party and customer-developed name databases.

Matching Features

Search parameters can use given value lists for input selection. Either as single or multiple values for matching.

Matching Features

Users can define similarities between values from a multiple-choice list. For example, users could specify hair color as “brown is 75% similar to light brown”..

Matching Features

ELISE ID Platform results are transparent. For each result item, users can therefore see precisely how the score was built up. Results include highlighting information and details on score calculation, so it is immediately clear why a result item is seen as a full or partial match.

Matching Features

ELISE ID Platform can use biometric or other image information to identify people stored in the database. This capability is provided through third-party algorithms. These algorithms can be quickly added, replaced, or updated at any time during implementation or production. As a result, there is no downtime.

ELISE ID Platform applications, services, and insights

For more information about the ELISE ID Platform, or to learn more about our out-of-the-box Identity and Security solutions, contact WCC.

Success Stories

eu-LISA

(Europe)

European Union

• The EU needed a central database for all visa applications in order to combat ‘visa shopping’.
• The EU VIS system uses WCC ELISE ID Platform with biographic matching and multi-cultural name matching capabilities.
• EU VIS is the most stable and highest-performing of all EU systems.
For more details, including the use of UMF to achieve EU interoperability, please visit this blog.

Instituto Nacional de Migración

(Mexico)

Instituto Nacional de Migración

• The large flow of immigrants crossing the Guatemalan border with Mexico called for border control and an Identity Management System.
• The Identity Management System uses ELISE ID Platform with integrated biographic data and biometrics.
• The system gives insights into migratory routes, records multiple encounters of the same person, and can also assist food allocation using biometric data.

UNHCR

(United Nations)

UNHCR

• The UN lacked insight into the movement of refugees between camps. Identity and rations fraud were major challenges.
• UNHCR’s Identity Management System uses WCC ELISE ID Platform with integrated biographical data and biometrics.
• UNHCR can now monitor refugee migration, and refugee families are reunited more quickly. ID and food ration card fraud has decreased significantly.