Database Reference
In-Depth Information
effectiveness of even routinely occurring transactions. This is because most end-user application's
effectiveness and performance can be enhanced by transforming it from a
bare
transaction to a
transaction
clothed
by a surround of a context formed as an aggregate of all relevant decision pat-
terns utilized in the past.
The decision patterns contributing to a transaction's context include the following:
◾
Characteristic and sundry details associated with the transaction under consideration
◾
Profiles of similar or proximate transactions in the immediately prior week or month or
6-months or last year or last season
◾
Profiles of similar or proximate transactions in same or adjacent or other geographical regions
◾
Profiles of similar or proximate transactions in same or adjacent or other product groups or
customer groups
To generate the context, the relevant decision patterns can either be discerned or discovered by
mining the relevant pools or streams of primarily the transaction data. Or they could be aug-
mented or substituted by conjecturing or formulating decision patterns that explain the existence
of these characteristic pattern(s) (in the pools or streams of primarily the transaction data). In the
next subsection, we look at function-specific decision patterns with a particular focus on financial
decision patterns.
Thus, generation of context itself is critically dependent on employing big data and
mobilized applications, which in turn needs cloud computing as a prerequisite.
14.9.1.1 Concept of Patterns
he concept of patterns used in this topic originated from the area of real architecture. Alexander
gathered architectural knowledge and best practices regarding building structures in a pattern
format. This knowledge was obtained from years of practical experience. A pattern according to
Alexander is structured text that follows a well-defined format and captures nuggets of advice on
how to deal with recurring problems in a specific domain. It advises the architect on how to create
building architectures, defines the important design decisions, and covers limitations to consider.
Patterns can be very generic documents, but may also include concrete measurements and plans.
Their application to a certain problem is, however, always a manual task that is performed by the
architect. Therefore, each application of a pattern will result in a differently looking building,
but all applications of the pattern will share a common set of desired properties. For instance,
there are patterns describing how eating tables should be sized so that people can move around
the table freely, get seated comfortably, find enough room for plates and food, while still being
able to communicate and talk during meals without feeling too distant from people seated across
the table. While the properties of the table are easy to enforce once concrete distances and sizes
are specified, they are extremely hard to determine theoretically or by pure computation using a
building's blueprint.
In building architecture, pattern-based descriptions of best practices and design decisions
proved especially useful, because many desirable properties of houses, public environments, cit-
ies, streets, etc. are not formally measurable. They are perceived by humans, and thus, cannot be
