Fraud Detection

plaid的payment risk就是这一类

我想到的metrics

• Offline

  • confusion matrix: recall
  • NE

• Online

  • number of fraud transaction
  • loss from fraud, per day, or per million transaction

看这里 [[TP, FN, TOC, etc.....]]

Precision

• Precision is calculated by dividing the true positives by anything that was predicted as a positive.

Recall

• Recall (or True Positive Rate) is calculated by dividing the true positives by anything that should have been predicted as positive.

Accuracy

• 就是正确率

Accuracy = (TP + TN) / Total Number

False Positive

• 也就是False decline, because there are reasons to suspect fraud
• Pain point for financial institution which stand to lose out on customer loyalty, when company incorrectly declines transaction

2015年，15% all carholders had at least 1 transaction INcorrectly declined in the previous year, which represented an annual revenue loss totaling nearly $118 billion

What’s more, 39% 被decline的人，gave up their card, after falsely declined

Overall, lost revenue and diminished customer loyaty

False Positive and False Negative will both loss money