|
|
|
|
Data Mart
Data warehouses can become enormous with hundreds of gigabytes
of transactions. As a result, subsets,
known as "data marts," are often created for just one department
or product line. Data Warehouse
combines databases across an entire enterprise. However, Data
Marts are usually smaller and focus on a
particular subject or department or product line.
Following are the common techniques through which a data
warehouse can be used.
11.1 Online Analytical Processing (OLAP)
Decision support software that allows the user to quickly
analyze information that has been
summarized into multidimensional views and hierarchies. The term
online refers to the interactive
querying facility provided to the user to minimize response
time. It enables users to drill down into
large volume of data in order to provide desired information,
such as isolating the products that are
more volatile from sales data. OLAP summarizes transactions into
multidimensional user defined views.
11.2 Data Mining
Data mining is also known as Knowledge-Discovery in Databases
(KDD). Put simply it is the
processing of the data warehouse. It is a process of
automatically searching large volumes of data for
patterns. The purpose is to uncover patterns and relationships
contained within the business activity
and history and predict future behavior. Data mining has become
an important part of customer
relationship management (CRM).
The data mining procedure involves following steps
•
Exploration – includes
data preparation which may involve filtering data and data transformations,
selecting subsets of records.
•
Model building and
validation – involves the use of various models for predictive performance
(i.e.,
explaining the variability in question and producing stable
results across samples). Each model
contains various patterns of queries used to discover new
patterns and relations in the data.
•
Deployment – That final
stage involves using the model selected as best in the previous stage and
applying it to new data in order to generate predictions or
estimates of the expected outcome.
Example of Data Mining
Consider a retail sales department. Data mining system may infer
from routine transactions that
customers take interests in buying trousers of a particular kind
in a particular season. Hence, it can
make a correlation between the customer and his buying habits by
using the frequency of his/her
purchases. The marketing department will look at this
information and may forecast a possible clientele
for matching shirts. The sales department may start a
departmental campaign to sell the shirts to buyers
of trousers through direct mail, electronic or otherwise. In
this case, the data mining system generated
predictions or estimates about the customer that was previously
unknown to the company.
Concept of Models Used in Decision Support System (DSS)
“A model is an abstract representation that illustrates the
components or relationships of a
phenomenon.”
38
Models are prepared so as to formulate ideas about the problem
solutions that is allowing the managers to
evaluate alternative solutions available for a problem in hand.
11.3 Types of Models Used in DSS
•
Physical Models
•
Narrative Models
•
Graphic Models
•
Mathematical Models
11.3.1 Physical Models
•
Physical models are
three dimensional representation of an entity (Object / Process). Physical
models
used in the business world include scale models of shopping
centres and prototypes of new
automobiles.
The physical model serves a purpose that cannot be fulfilled by
the real thing, e.g. it is much less expensive
for shopping centre investors and automakers to make changes in
the designs of their physical models
than to the final product themselves.
11.3.2 Narrative Models
The spoken and written description of an entity as Narrative
model is used daily by managers and
surprisingly, these are seldom recognized as models.
For instance
All business communications are narrative models
11.3.3 Graphic Models
These models represent the entity in the form of graphs or
pictorial presentations. It represents its entity
with an abstraction of lines, symbols or shapes. Graphic models
are used in business to communicate
information. Many company’s annual reports to their stockholders
contain colourful graphs to convey
the financial condition of the firm.
For Instance
Bar graphs of frequently asked questions with number of times
they are asked.
11.3.4 Mathematical Models
They represent Equations / Formulae representing relationship
between two or more factors related to
each other in a defined manner.
Types of Mathematical Models
Mathematical models can further be classified as follows, based
on
•
Influence of time –
whether the event is time dependant or related
•
Degree of certainty –
the probabilities of occurrence of an event
39
•
Level of optimization –
the perfection in solution the model will achieve.
Hence use of right model in decision support software is
critical to the proper functionality of the system.
Group DSS
When people responsible for decision making are geographically
dispersed or are not available at a place at
the same time, GDSS is used for quick and efficient decision
making. GDSS is characterized by being
used by a group of people at the same time to support decision
making. People use a common
computer or network, and collaborate simultaneously.
An electronic meeting system (EMS) is a type of computer
software that facilitates group decision-making
within an organization. The concept of EMS is quite similar to
chat rooms, where both restricted or
unrestricted access can be provided to a user/member.
DSS vs. GDSS
DSS can be extended to become a GDSS through
•
The addition of
communication capabilities
•
The ability to vote,
rank, rate etc
•
Greater system
reliability
11.4 Knowledge / Intelligent Systems
Before we proceed with defining these systems, first we should
have clear concept of Knowledge
Management. The set of processes developed in an organization to
create, gather, store, maintain and
apply the firm’s knowledge is called Knowledge Management. Hence
the systems that aid in the
creation and integration of new knowledge in the organization
are called knowledge systems.
There are two questions
Who are they built for?
This refers to defining the knowledge workers for whom the
knowledge system is being built. The term
refers to people who design products and services and create
knowledge for an organization. For instance
Engineers
Architects
Scientists
•
Knowledge systems are
specially designed in assisting these professionals in managing the knowledge in
an organization.
What are they built for?
Every knowledge system is built to maintain a specific form of
knowledge. Hence it needs to be defined
in the start, what the system would maintain. There are major
types of knowledge.
•
Explicit knowledge –
Structured internal knowledge e.g. product manuals, research reports, etc.
•
External knowledge of
competitors, products and markets
•
Tacit knowledge –
informal internal knowledge, which resides in the minds of the employees but
has not been documented in structured form.
Knowledge systems promote organizational learning by
identifying, capturing and distributing these forms
of knowledge
40
11.5 Knowledge Support Systems (KSS) / Intelligent Systems
These systems are used to automate the decision making process,
due to its high-level-problem-solving
support. KSS also has the ability to explain the line of
reasoning in reaching a particular solution, which
DSS does not have.
Intelligent Systems
Knowledge systems are also called intelligent systems. The
reason is that once knowledge system is up and
running, it can also enable non experts to perform tasks
previously done by experts. This amounts to
automation of decision making process i.e. system runs
independently of the person making decisions.
Artificial Intelligence
“Artificial intelligence is the ability of a machine to
replicate the human thought processes. The way humans
proceed to analyze a problem and find appropriate solutions,
similarly computers are geared up to follow
human logic to solve problems.”
These knowledge-based applications of artificial intelligence
have enhanced productivity in business,
science, engineering, and the military. With advances in the
last decade, today's expert systems clients can
choose from dozens of commercial software packages with
easy-to-use interfaces.
The most popular type of intelligent systems is the Expert
System.
Expert System
An expert system is a computer program that attempts to
represent the knowledge of human experts in
the form of Heuristics. It simulates the judgment and behaviour
of a human or an organization that has
expert knowledge and experience in a particular field.
Examples are
•
Medical diagnosis,
•
Equipment repair,
•
Investment analysis,
•
Financial, estate and
insurance planning,
•
Vehicle routing,
•
Contract bidding
Heuristics
Heuristic is the art and science of discovery and invention. The
word comes from the same Greek root
as "eureka", which means "I have found it". A heuristic is a way
of directing your attention fruitfully. It
relates to using a problem-solving technique, in which the most
appropriate solution is found by
alternative methods. This solution is selected at successive
stages of a program for use in the next step
of the program.
11.6 Components of an Expert System
There are four main components of Expert systems
41
•
User Interface: to
enable the manager to enter instructions and information into an expert system
to
receive information from it.
•
Knowledge Base: it is
the database of the expert system. It contains rules to express the logic of the
problem.
•
Inference engine: it is
the database management system of the expert system. It performs reasoning by
using the contents of the knowledge base.
•
Development engine – it
is used to create an expert system.
Neural Network
Hardware or software that attempt to emulate the processing
patterns of the biological brain. It is a device,
modeled after the human brain, in which several interconnected
elements process information
simultaneously, adapting and learning from past patterns.
Neural Network vs. Expert System
Expert systems seek to model a human expert’s way of solving
problems. They are highly specific to seeking
solutions. Neural networks do not model human intelligence. They
seek to put intelligence into the
hardware in the form of generalized capability to learn.
Fuzzy Logic
The word Fuzzy literally means vague, blurred, hazy, not clear.
Real life problems may not be solved by an
optimized solution. Hence allowance needs to be made for any
imperfections which may be faced while
finding a solution to a problem. Fuzzy logic is a form of
algebra employing a range of values from “true” to
“false” that is used in decision-making with imprecise data, as
in artificial intelligence systems. It is a rule
based technology that tolerates imprecision by using non
specific terms/ imprecise concepts like "slightly",
"quite" and "very". to solve problems. It is based on the
Possibility theory, which is a mathematical theory
for dealing with certain types of uncertainty and is an
alternative to probability theory.
Executive Support Systems (ESS)
This Computer Based Information System (CBIS) is used by senior
managers for strategic decision making.
The decisions at this level are non-routine and require judgment
and evaluation. They draw summarized
information from internal MIS and Decision Support Systems.
These systems deal with external influences
on an organization as well.
•
New Tax laws
•
Competitors
•
Acquisitions,
take-overs, spin offs etc.
They filter, compress and track critical data so as to reduce
time and effort required to obtain information
useful for executives. They are not designed to solve specific
problems. They are generalized to be capable
of dealing with changing problems. Since executives have little
contact with all levels of the organization,
ESS uses more graphical interface for quick decision making.
ESS vs. DSS
ESS implies more of a war room style graphical interface that
overlooks the entire enterprise. A decision
42
support system (DSS) typically provides a spreadsheet style
"what if?" analysis capability, often for only one
department or one product at time. |
|
|
|
|
|
|