You are on page 1of 3

Online Fraud

Detection
About the topic
 If you're selling or buying something online providing financial services or simply processing tons
of payments , you face fraud risks everyday..
Fraud is all around and every organization that is conducting business with customers exposes itself
to the risk of fraud

So detecting and preventing fraud is essential

 For payment fraud we have large number of interacting parties... In fact , about 10 years ago
we have credit cards and our box office in bank only .But now we have digital banks , card
networks such as visa or mastercard , digital wallet like apple pay or google pay , biometric
notifications , cryptocurrencies... And attack to all these has serious consequences
 As we know , Machine learning based systems are extremly good at finding patterns that humans
can't . It knows very nuanced distinctions between normal and fraudulent activities
 As we know , Machine learning based systems are extremly good at finding patterns that
humans can't . It knows very nuanced distinctions between normal and fraudulent
activities .
1. The first step is understanding what is normal to identify what deviates from that
normal so training on the historical data : past transactions both fraudulent and not.
Each transaction has different attributes :
* the exact time and date of transaction
* Purchase amount
* accountID / transactionID
…...
1/ import both the transaction data and fraud data in Python
2/ treat tha data : convert the type of transaction times to hhmmss , change the type of
transaction dates to string , sort the datasets by accountID date and time transactions
, remove duplicates from both files

You might also like