Expert
Systems
The expert systems are just computer
software. “They are programs that achieve expert level competence in solving
problems in particular task area by use of knowledge base about that particular
task area”(1).
By other words, the expert system is a type of computer system that can perform
a task of emulates the decision making that otherwise can be done by a human
expert. However, the expert system will do such a task with more efficient
abilities and sufficient capability. This type of systems was designed to deal
with very complicated problems using the rules of IF-Then instead of following
the computer conventional ways.
Expert
systems are known as artificial intelligent (AI) programs. These programs are
widely spread in different sectors and fields such medical, business, and agriculture.
In agriculture, for example, we can recognize CLIPS (C Language Integrated
Production System). It is the most expert system been used in this
field. It was developed in 1985 at NASA and known today as fast, efficient and
free expert system. This expert system in agriculture has had provided a
recognizable many advantages to the agricultural field:
-
has the ability to imitate human thought and reasoning
-
Help to increase the production of corps.
-
has the ability to handle uncertain information.
-
help farmers to take single point decision
Neural Networks
Progressively, Neural networks
becoming more and more prevalent in business world. Nowadays, many businesses
are spending more on researches that focusing on neural network and data mining
to find appropriate solutions that make the numerous business data readily. It
is became obvious that such technologies and methodologies will help any
company ( dealing with a wide range of business data) forecasting, modeling,
clustering, and classification its data. As a result, decision making will be
able to generate affective, useful and successful business plans. However, such
results might need an expert to read them right. Thus what are neural networks?
Neural networks are a functional
approximation tools that study the relationship between independent and dependent
variables, “much like regression or other more traditional applications”(2).
The only difference between the neural networks and statistical approaches “is
that neural networks make no assumptions about the statistical distribution or
properties of the data”(2). Additionally, the neural networks are
classifies to several types based on its purpose, architecture, and its
learning algorithm. There are many of those types such as:
1- Multilayered feed forward
neural networks
2- Hopfield neural networks
3- Self-organizing neural networks
Some studies asserted that these
are the three main models of artificial neural networks (ANN).
The business applications that use
these type of ANNs are very helpful in many ways. They help identifying the
customers who are, definitely, expected to buy and use the product provided.
Such applications also had ability to predict customers’ behavior. Computer
operates usually follow a sequential linear processing techniques and “apply
some formulas, decision rules, and algorithms”(3) to provide results
that decision makers can use. Differently, the ANNs improve their own instructions
and generate better results. There for, this type of applications can help to
determine when a customer is about to switch to use a competitor’s product or service.
The ANNs have had illustrated a
great success in terms of sales forecasting. This fact has been confirmed due
to ANNs’ abilities to consider multiple variables at a time.
Trading and financial
forecasting also was benefited from ANNs applications. They been used in
solving problems related to “derivative securities pricing and hedging, futures
price forecasting, exchange rate forecasting and stock performance and
selection prediction”(2).
Likewise, the insurance industry can
benefit from such ANNs applications as well. Since Neural networks applications
can multiply different variables at a time, the policy holders can be divided
int many segments based on their behaviors for example. The ANNs provided
results that help to determine effective premium pricing plans.
The field is increasingly grown,
and the business applications that use the technologies of ANNs are expanding
as well. It is hard to list or enumerate all of them; but we can recognize the
following as most popolar such as: Airline
security control.
Investment management and risk control.
Prediction of thrift failures.
Prediction of stock price index(3).
Sources: