Information Technology Reference
In-Depth Information
improve the services, mere classification rules are not sucient. In this paper, we
propose to use classification rules to build a new strategy of action based on their
condition features in order to get a desired effect on their decision feature. Going
back to the bank example, the strategy of action would consist of modifying some
condition features in order to improve our understanding of customers behavior
and then improve the services. E-action rules are useful in many other fields,
including medical diagnosis. In medical diagnosis, classification rules can explain
the relationships between symptoms and sickness and in predicting the diagnosis
of a new patient. E-action rules are useful in providing a hint to a doctor what
symptoms have to be modified in order to recover a certain group of patients
from a given illness.
Action Tree algorithm is presented for generating E-action rules and it is
implemented as System DEAR 2 . 2. The algorithm follows a top-down strategy
that searches for a solution in a part of the search space. It is seeking at each
stage for a stable attribute that has a least number of values. Then, the set
of rules is split recursively using that attribute. When all stable attributes are
processed, the final subsets are split further based on a decision attribute. This
strategy generates an action tree which is used to construct E-action rules from
the leaf nodes of the same parent.
9.2
Information System and E-Action Rules
An information system is used for representing knowledge. Its definition, pre-
sented here, is due to Pawlak [12].
By an information system we mean a pair S =( U, A ), where:
U is a nonempty, finite set of objects,
A is a nonempty, finite set of attributes i.e. a : U
−→
V a is a function for any
a
A ,where V a is called the domain of a .
Elements of U are called objects. In this paper, for the purpose of clarity,
objects are interpreted as customers. Attributes are interpreted as features such
as, offers made by a bank, characteristic conditions etc.
We consider a special case of information systems called decision tables [12].
In any decision table together with the set of attributes a partition of that
set into conditions and decisions is given. Additionally, we assume that the set
of conditions is partitioned into stable conditions and flexible conditions. For
simplicity reason, we assume that there is only one decision attribute. Date of
birth is an example of a stable attribute. The interest rate on any customer
account is an example of a flexible attribute as the bank can adjust rates. We
adopt the following definition of a decision table:
By a decision table we mean any information system S =( U, A St
A Fl ∪{
d
}
),
where d
A Fl is a distinguished attribute called the decision. The elements
of A St are called stable conditions, whereas the elements of A Fl are called flexible
conditions.
A St
 
Search WWH ::




Custom Search