Fraud detection evaluation formula

Part 1: Scoring for each criterion

To create a formula that helps assess the risk level of customer fraud (low, medium, high), we can use the following criteria:

1. Using Proxy/VPN or Web Proxy to Place Orders
   1. Check: whether the IP address used to place the order is a proxy/VPN or identified as a web proxy connection.
   2. Scoring Criteria:
   • Not using a proxy/VPN: 0 points
   • Using a proxy/VPN or web proxy: 3 points
2. Special Character in Order Information
   1. Check: If the customer's name and shipping address contains special characters (@, #, $, %,…)
   2. Scoring:
      • No special character: 0 point

      • Contains special characters: 2 points

        Distance between order IP and Shipping address
   3. Check the distance between the IP and shipping address
   4. Scoring:
      • =<20km: 0 point
      • 20<distance=<50km: 1 point
      • 50<distance=<200km: 2 points
      • >200km: 3 points

Part 2: Aggregating Scores and Assessing Fraud Risk Level

1. Total Score: Sum the points from all the criteria above (Total score = Proxy/VPN + Special characters points + Distance points)
2. Fraud Risk Level:
   1. Low: Total score from 0 to 2
   2. Medium: Total score from 3 to 5
   3. High: Total score of 6 or more
3. Examples:

• Customer A:

• Use VPN (3 points)
• No special characters (0 points)
• Distance between IP and shipping address: 100 km (2 points)

=> Total score: 3 + 0 + 2 = 5 → Medium

• Customer B:

• Use VPN (3 points)
• Contains special characters (2 points)
• Distance between IP and shipping address: 300 km (3 points)

=> Total score: 3 + 2 + 3 = 8 → High