A Data Mining Based Fraud Detection Hybrid Algorithm in E-bank

Many banks now provide the online bank service, allowing customer to do transaction online. However, it also boosts the telecom fraud especially for the credit card fraud. In past several years, some studies have worked on some data mining-based method to detect online transaction fraud. However, mo...

Full description

Saved in:
Bibliographic Details
Published in2020 International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE) pp. 44 - 47
Main Author Song, Zijian
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2020
Subjects
Artificial intelligence
Big Data
Conferences
Data Mining
Feature engineering
Fraud detection
Hybrid algorithm
Indexes
Internet of Things
Online AccessGet full text
DOI10.1109/ICBAIE49996.2020.00016

Cover

More Information
Summary:Many banks now provide the online bank service, allowing customer to do transaction online. However, it also boosts the telecom fraud especially for the credit card fraud. In past several years, some studies have worked on some data mining-based method to detect online transaction fraud. However, most of studies fitted with little data, small amount of features and just single model classifier. Moreover, the model classifiers used by them are usually too weak to fit with large set of data. To handle this problem, in this paper, we proposed a hybrid data mining-based algorithm to detect fraud. The dataset is IEEE-CIS Fraud dataset provided by E-banks and it is big enough to train a good classifier. In addition, proper feature engineering is done to achieve a better accuracy. In the experiment section, we compare our hybrid method with the other classical method with single model to show the power of our method.
DOI:10.1109/ICBAIE49996.2020.00016