PENGGUNAAN DAN OPTIMASI ALGORITME DECISION TREE PADA SISTEM DETEKSI FRAUD
The use of mobile money technology in financial transaction services provides various conveniences for its users. Apart from its usefulness, mobile money, like any other means of payment, is also vulnerable to various abuses. One form of abuse that is often encountered is fraud. Fraud not only ca...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/62118 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The use of mobile money technology in financial transaction services provides various
conveniences for its users. Apart from its usefulness, mobile money, like any other
means of payment, is also vulnerable to various abuses. One form of abuse that is often
encountered is fraud. Fraud not only causes financial loss to users and providers of
mobile money services, but also damages the reputation of service providers and the
entire industry. To deal with fraud that occurs in any transaction activity, an intelligent
system is developed using a model built from machine learning algorithms. In this final
paper research, fraud detection system was developed using a model based on decision
tree algorithm its variations. Implementation is carried out through five stages
following the CRISP-DM method. To improve the performance of the model,
optimization is done by finding and selecting the best combination of hyperparameter
values for each algorithm. From the tests carried out, it was found that the performance
of the model reached metric accuracy values of 99.91%, precision of 99.9%, recall of
72.19%, F1 score of 81.93%, ROC AUC of 86.1%, and FPR of 0.0002%. |
|---|