Credit Card Fraud Detection using Anomaly Detection

MJ·2021년 4월 24일
0

AnomalyDetection

목록 보기
1/1

I. Introduction

  • Detecting a fraud is a tricky computational piece of work
    • very difficult to choose parameters.
    • The success of fraud detection depends upon cluster and classification of parameters

II. Anomaly Detection

  • Anomalies are those values and patterns that do not belong to the normal behavior
    • A1 and A2 are Anomalies

III. Different aspects of anomaly detection

  • Description of Input Data:
  • Data labels
    • supervised data
    • semi supervied data
    • unsupervised data
  • Types of anomaly
    • Point Anomaly
    • Contextual Anomaly
    • Collective Anomaly
  • Consequence of Anomaly Detection
    • Score
      • each instance is assigned a score
      • depending on degree, finds out that anomalous instance
      • use cut off threshold to select anomalies
    • Labels
      • labeling is used for normal and anomalies
      • assign to each data instance
      • uses a domain specific threshold to choose anomalies
      • use binary labels to examine instance may assign 0 as anomalies and 1 as normal
profile
Master of Science in Statistics

1개의 댓글

comment-user-thumbnail
2023년 5월 18일

I think you will find this site improve credit history useful. You can use this service, users can improve credit history. Also this service is legitimate, you can buy an apartment, take out insurance or buy a car as you will have to provide the results of the credit check.

답글 달기