Data Mining Implementation
Data Mining Implementation
Conclusion
Based on the case studies that have been carried out, starting from the stage of
literature study, observational studies, system design and implementation, the results obtained
can be drawn the following conclusions:
1) The clustering method with a divisive hierarchical algorithm can be used to group customers
to make the company's business intelligence competitive.
2) Information from a customer's transaction frequency can be used to build a system that can
transform customer data into information that is useful for carrying out competitive intelligent
company business processes.
3) The program is only designed for one user, namely the marketing manager and customer
service, so that other unauthorized users will not be able to access this program because the
username and password as valid program access rights are only designed for one user and no
facilities are provided for adding access rights.
4) The program will carry out the updating process automatically if there is a change in the
database, especially in the transaction table and customer table for existing data values in the
transaction frequency table and also the cluster process result tables and percentage graphs.
5) To form a cluster, a central point is needed which can be found from all the data in the
transaction frequency table by calculating the mean value or median value calculation method.
6) Applications can still run well when run on three different web browsers, namely Internet
Explorer, Mozilla Firefox and Flock web browsers.
7) The results of this application can be used as decision support by managers for their
customers. For example, decision support to increase promotions to some customers who are
in less active and moderate clusters or decisions to provide more exclusive facilities or give
bonuses or discounts to customers in active clusters.