DiscussionObviously, many methods a

Discussion
Obviously, many methods are introduced for recruitment
in management science. Non-use of information technology
in this field and using the idea of knowledge hidden in the
database of current staff performance creates new issues in
this field. The main objective of this study is to prevent the
improper selection of employees by using statistical and
commonly known methods.

The purpose of this research was to search the database
for employee performance by using data mining methods
in order to discover patterns of employee efficiency and
effectiveness and to convert this information into useful
knowledge for the organisation. After discovering these
patterns, the rules related to factors affecting performance
were extracted and provided to managers. This will help
them select the best employees.

According to the comments of bank experts about rules,
some conclusions derived are as follows:
• Deleting of target variables such as ‘performance
assessment’ demonstrates that system performance
evaluation is not implemented properly.
• According to derived rules, multiple-choice tests
conducted are not standard and not appropriate criteria
for use in an interview session.
• Age, gender and marital status of employees have no
effect on their upgrade’s status.
• Exam courses such as literature, education, mathematics
and general knowledge have no effect on performance
and promoting those selected.
• Some upgrades have been made when an employee
submits a degree after employment.
• Employees’ university of choice has an effect on the
interview.
• Employees’ major field of study has no effect on
employment.
• With regard to the accuracy of extracted rules, higher
knowledge represents superior candidates.
• The main reason for promoting people in the Northern
areas is related to the extent of these areas and the fact
that posts are emptied earlier in these areas.
Although this information is openly available to managers,
documents can now be submitted to check this with
confidence.

Finally, it is obvious that knowledge cannot necessarily be
obtained from extracted rules. Other research should be done
to provide additional knowledge.

The following recommendations can be submitted to
improve the study:
• With regard to the removal of test variables, it is essential
to change the type of selected course and its content.
• It is better to change scoring methods for performance
evaluation and training tools should be used for
completing information.