Data Mining On Retail Banking Simulation Model To Devise Productivity Improvement Strategies
Retailing banking is service-oriented business. Customers visit different branches to procure services such as transaction inquiry, process bank account, etc. The scenario is operators of various positions of a branch have to render counter services and meet certain customer service level. Cus...
Saved in:
Main Author: | |
---|---|
Format: | Monograph |
Language: | English |
Published: |
Universiti Sains Malaysia
2019
|
Subjects: | |
Online Access: | http://eprints.usm.my/58291/1/Data%20Mining%20On%20Retail%20Banking%20Simulation%20Model%20To%20Devise%20Productivity%20Improvement%20Strategies.pdf http://eprints.usm.my/58291/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Retailing banking is service-oriented business. Customers visit different branches
to procure services such as transaction inquiry, process bank account, etc. The scenario is
operators of various positions of a branch have to render counter services and meet certain
customer service level. Customers arrive at the counter at different patterns. Information
can be collected, such as customer waiting time, operator on service time, number of tickets
received at particular time. Such information allows data mining to be performed to
discover patterns that could segregate the type of services. The study aims to demonstrate
such possibility through computer simulation. Specifically, investigation is made for the
data mining to determine the clustering and classification of the branches and operator
productivity respectively. The research methodology involves seven steps. First, business
understanding is performed to gain insight into the motive of initiating data mining exercise.
Two levels, macro and micro levels were defined to differentiate inter and intra-branches
comparisons. Second, a computer simulation would be constructed in WITNESS Horizon
V.21, largely based on the description of a real branch. Different scenarios were built
reflecting operating behaviors of different branches. The simulation stores data (about six
thousands of records) in a database. The following steps involve different data selection
processing strategies (selection, cleaning, transformation). Next is the data mining,
primarily using Python Orange V.3.20. Last step will be pattern evaluation to develop
suitable productivity improvement strategies. In this research, a variety of data mining tools
have been deployed and multiple insights were generated. Notably, branches adopted lean
management have shown improved in general productivity. Operator performances were
able to differentiate based on years of experience. This research provides opportunities for
researchers to examine the productivity of branches with other data mining techniques. It
also helps bankers to focus on the right areas to be improved and increase the ability in
decision making. There are two main limitations of the research. First, the simulation model
does not capture the whole intricacy of retail banking front-end operations. Second,
software limitations hindered more complex data mining tasks to be performed. |
---|