CREDIT CARD INTERNAL FRAUD DETECTION ON ONLINE TRANSACTION DATA USING HIDDEN MARKOV MODEL
The rapid progress in the technology of electronic commerce has led to increased use of credit cards. This happens because credit cards are an easy payment method for online shopping transactions. With the increase in the use of credit cards, cases of credit card fraud have also increased. If in...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/36178 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The rapid progress in the technology of electronic commerce has led to increased
use of credit cards. This happens because credit cards are an easy payment method
for online shopping transactions. With the increase in the use of credit cards, cases
of credit card fraud have also increased.
If information about credit card transactions is an observed process and the factors
that cause the magnitude of credit card transactions are not directly observed
(hidden) and assumed to form a Markov chain, then credit card transactions can
be modeled by the Hidden Markov Models (HMM).
In this study, will be discussed the process of traning a model that records the normal
behavior of a credit card holder until the process of detecting new transactions on a
credit card.
After detecting new cret card transactions, an optimal threshold value will be determined
for a new transaction to be categorized as fraud transaction. Several experiments
with different amounts of hidden state will be presented to show the effectiveness
of model in detecting fraud transaction and the effectiveness of model
against the level of accuracy. |
|---|