Fraud detection systems and methods
Systems and methods disclosed herein related to a fraud detection system that receives data from multiple sources and normalizes the data to create transaction data using overlapping data from the multiple sources. Fraud analysis is performed using a probabilistic approach to alert generation, wherein alerts are only sent to customers when the probability of fraud exceeds a threshold. As a result, false positives are avoided.
1. A computer-implemented method for removing personally identifiable data comprising:
receiving, by a computer of a fraud detection system, financial transaction data associated with one or more financial transactions from a server of a first financial institution, wherein the server of the first financial institution records the financial transaction data in a non-transitory memory;
identifying, by the computer of the fraud detection system, personally identifiable data within the financial transaction data received from the first financial institution;
generating, by the computer of the fraud detection system, a first set of hashed data of the personally identifiable data by applying a hashing algorithm using a first hash key to the personally identifiable data;
generating, by the computer of the fraud detection system, a first reverse hashing map by creating a table including information used for converting the first set of hashed data back into the personally identifiable data received from the first financial institution;
transmitting, by the computer of the fraud detection system, the first reverse hashing map to the server of the first financial institution;
deleting, by the computer of the fraud detection system, the first reverse hashing map and the first hash key;
performing, by the computer of the fraud detection system, fraud analysis using the first set of hashed data;
generating, by the computer of the fraud detection system, a report containing the likelihood one or more transactions recorded in the financial transaction data are associated with fraud, wherein the report contains first set of hashed data instead of the personally identifiable data;
transmitting, by the computer of the fraud detection system, the report to the server of the first financial institution; and
retrieving, by the server of the first financial institution, the personally identifiable data from the report using the first reverse hashing map and the first set of hashed data.
2. The method according to claim 1 , further comprising determining, by the computer of the fraud detection system, a likelihood one or more financial transactions in the financial transaction data associated with the first financial institution is associated with fraud based upon a fraud detection analysis of the first set of hashed data and the financial transaction data from the first financial institution.
3. The method of claim 1 , further comprising receiving, by the computer of the fraud detection system, financial transaction data associated with one or more financial transactions from a server of a second financial institution, wherein the server of the second financial institution records the financial transaction data in a non-transitory memory.
4. The method according to claim 3 , further comprising: applying, by the computer of the fraud detection system, a hashing algorithm using a second hash key to the personally identifiable data in the financial transaction data received from the second financial institution, thereby generating a second set of hashed data representing the personally identifiable information from the second financial institution;
generating, by the computer of the fraud detection system, a second reverse hashing map, wherein the second reverse hashing map explains how to restore the personally identifiable information of the second financial institution from the code generated by the hashing algorithm;
transmitting, by the computer, the second reverse hashing map to the second financial institution; and
deleting, by the computer of the fraud detection system, the second reverse hashing map and the second hash key.
5. The method according to claim 4 , further comprising determining, by the computer of the fraud detection system, a likelihood one or more transactions associated with the financial transaction data of the second financial institution is associated with fraud based upon a fraud detection analysis of the second set of hashed data and the financial transaction data from the second financial institution.
6. The method according to claim 4 , further comprising automatically selecting, by the computer of the fraud detection system, the second hash key for the second financial institution.
7. The method according to claim 1 , wherein the personally identifiable data is represented by a collection of numbers and letters after applying the hashing algorithm.
8. The method according to claim 1 , wherein the hashing algorithm is applied to obfuscate the personally identifiable data.
9. The method according to claim 1 , wherein the hashing algorithm is secure hash algorithm 1 (SHA-1), secure hash algorithm 2 (SHA-2), or secure hash algorithm 3 (SHA-3).
10. The method according to claim 1 , further comprising periodically updating, by the computer of the fraud detection system, the hash key.
11. The method according to claim 1 , further comprising applying, by the computer of the fraud detection system, a same hash key across all financial institutions, such that the computer identifies a same customer with accounts at multiple financial institutions.
12. The method according to claim 1 , further comprising merging, by the computer of the fraud detection system, additional information describing a same transaction from different financial institutions.