Credit Card Anomaly Detection in Graph Database Using Controlled Mutual Information and Fuzzy Decision Rule

  • Unique Paper ID: 151157
  • Volume: 7
  • Issue: 12
  • PageNo: 81-87
  • Abstract:
  • In recent years, e-commerce grows an important credential for global trade, the observation of anomaly detection which identifies the abnormal behavior in fraud detection of credit card transactions has turned into an interesting field of research. This paper emphases on automatic credit card fraudulent transaction detection with the graph related features. This work does two important tasks they are determining significant features using controlled mutual information and vagueness of credit card information handing is achieved by using fuzzy decision rule. The main objective of the feature subset selection is to increase the maximization of relevancy and to reduce the redundancy among attributes to attributes. The proposed controlled mutual information uses the attribute to attribute relationship along with the class label in a supervised manner to improve the relevancy rate of selected feature sets and the class attribute. The proposed fuzzy decision rule is used to infer the knowledge about pattern of credit card dataset and establish a classification model which can effectively handle the inconsistency in determining anomalous which is known as fraudulent transaction even in the imbalance dataset. From the performance analysis it is observed that the proposed model produces better results in credit card fraudulent detection.

Related Articles