 |
|
|
 |
What's New?
Revamped Memory Management - VBase Explained
VBase is the new, persistent object repository for the Visual Clarity Server. It leverages an object-relational mapping
layer to facilitate the storage and retrieval of the various types of VisuaLinks objects in an enterprise relational
database management system. VBase is designed to remove the memory-bound limitations from VisuaLinks implementations.
Ultimately, intermediate processing results for large-scale operations (like Network Miner) will be stored in VBase.
VBase can currently be configured to use Oracle 9i, SQL Server 2000, Sybase, PostgreSQL, MySQL, and Access as the physical
data store at installation time. The backup and restore capability of VBase will facilitate the easy migration from one
RDBMS to another should this be necessary. It will also play a role in the upgrade path from one version of VisuaLinks to the next.
VBase provides transactional data integrity to the various object types managed by the Visual Clarity Server such as Users,
Groups, Models, Alias Names, Client Properties, Exclude Lists, Hit Lists, Search Lists, and Mail, to name a few. It also
provides an object caching capability which equates to faster processing of user requests by the Visual Clarity Server.
In addition, a powerful new capability has been introduced that allows a user to save Information Sets from the current
working View to a folder in VBase. These Information Sets can be viewed and maintained by multiple users simultaneously.
One or more Information Sets (and/or models) can then be selected for processing from any of the VisuaLinks service dialogs
(e.g., Network Miner, Name Matcher, Duplicate Detector, etc.).
Using a new object edit interface, you can add, delete and edit objects and associations - including their respective
data elements. Change history to these data elements is automatically maintained as changes are made. This powerful new
feature means that VisuaLinks can not only be used to analyze your existing enterprise data via its models, but can also
be used to capture and analyze your manually entered case information.
|
|
|
 |
Here are a few tips and tricks to help you use VisuaLinks quickly and more efficiently.
|
|
|
|
|
 |
Database Query--Basic Query Mode
The Database Query now includes a "Basic" query mode that lets you execute a simple query that uses a single data source.
The basic query interface is a "Query by Example" interface. A Query by Example interface is a quick and easy way to query
where you don't necessarily have to know the location or names of the attributes you wish to query. As you select an object,
the attributes available for that object are shown. To query a value, simply enter a value into a field to use as a query
criterion. By default, the query selects only records in the database that match all the values entered.
To open the Database Query window, choose the Services Panel and then click Database Query  . From the Basic tab, select a
model and one or more types. The attributes available for each of the selected objects and/or associations are displayed in
the window at the right.
For each attribute you wish to search, enter a value and an optional operator. The  icon retrieves a handful of records
from the database and displays values for that attribute in the drop-down. This allows you to review the format of the
data in the attribute. It isn't really intended as a source for criterion selection.
The  icon allows you to select operators, such as Equal To and Greater Than.
Equal To is the default operator.
The Must Match All option is normally selected and requires that all attribute conditions be matched in order to return a record
from the database. Uncheck the Must Match All checkbox if you only want to require that at least one attribute condition to
be met in results returned.
Advanced Query - New Features
The Advanced Query interface looks similar to the current Database Query. The locations of some icons have changed, but the
basic operation remains the same. There are, however, two new features in the VisuaLinks 3.0 Database Query:
 Compare Columns and  Database Functions.
Applying Database Functions to Attributes
The icon displays the Database Functions window. From the Database Functions dialog you can select a database function to apply against the query results for the attribute. The function can also be applied to query result groups.
For example, the query below finds the youngest suspect for each eye color.
The results of this query are shown below:
Comparing Attribute Columns
The Compare Columns feature fills in the Query Value field with the selected object's attribute names allowing you to compare the values of one attribute with values of another attribute. This is a "sticky" button in that clicking the icon activates it and clicking it again deactivates it. An example is shown below:
The results of this query are shown below. You can see that the DATE_END and DATE_START attributes match.
|
|
 |
This month's Link Chart is derived from the recent success of the insurance Fraud Detection System (FDS) developed for the
Korean Financial Supervisory Services (FSS) by our partner, SolutionMart (www.solutionmart.net).
The chart is based on claim data derived from over 40 different insurance companies using a standard XML representation
for the core-data elements.
The Fraud Indicator (FI) database used by FSS is regularly updated based on hundreds of different criteria such
as number of accidents claimed, total payments made, date proximities from contract to accident, among many others.
An FI value is applied to all subjects within the database. Additionally, the data contains the names of the
claimants, injured, and other parties to the claim. The following shows a basic representation of the
underlying data that is being analyzed.
Using the Network Miner service, the analysts apply the SUM option to the FI (Fraud Indicator - a value assigned
to each subject to indicate the level of fraud interest). The results are shown below and sorted by this amount.
The top entry shows a network with an FI value of 3540. This network is selected for further review.
Click to Enlarge
Once the selected network is loaded, the following display is presented. Our experience tells us that the most
likely crime pattern consists of groups with some type of circles (closed network of subjects) in which there
are mixtures of accidents occurring on different dates and with different vehicles involved.
In the diagram shown below, the red lines indicate that the subjects ride in different vehicles and the blue
line indicates that they ride in the same ones. Also present are some heavily connected entities,
referred to as articulation points - because they hold the network(s) together and are considered critical.
Click to Enlarge
Within this diagram, the data has been arranged to better observe the "circle" patterns where loops, circles,
circuits, and closed connections are shown. The next diagram shows an area of the network (highlighted in red)
that is of interest and will be reviewed in more detail.
The next diagram reveals the degree of interaction among these subjects.
The names of the subjects have been blocked out and the FI number appears below the icon. The larger the number,
the more "suspicious" the subject is based on the contents of the fraud database. Additionally, in this sub-network,
the dates of the accidents appear on the linkages. A quick examination shows there are 10 unique dates occurring
roughly over a 12 month period - almost one a month.
The results of this analysis have lead to an active investigation into the fraudulent claims associated with these subjects.
|
|
|
 |
|
…you can adjust how the data view displays objects?
gives you a number of options for changing the layout of the icons and associations displayed in the View.
in the lower-right of the panel area. This floats the panels in a separate window and provides more space in the View area.
