Machine Learning
for Fraud Detection
Evertec is a full-service transaction processing company operating across 19 different countries in Latin America and the Caribbean. It processes over 2.1 billion transactions annually and employs more than 2600 people.
Source: Evertec.
CHALLENGE
Evertec uses a fraud detection solution for credit and debit transactions. However, their internal team was facing challenges related to the following aspects:
This challenge involved the security of the company's transactional management.
Evertec uses a fraud detection solution for credit and debit transactions. However, their internal team was facing challenges related to the following aspects:
- Accurately score the risk of credit and debit transactions to prevent fraud.
- Reduce the False Positive to True Positive ratio.
This challenge involved the security of the company's transactional management.
SOLUTION
We helped Evertec understand and curate their data (500 million records) by applying Data Science techniques. Complex Machine Learning detection models that outperformed the ones already in use were created.
This project shows an interesting case study in which Machine Learning applied to Big Data can have a noticeable impact on a company's turnover and therefore its profitability.
The graph shows the multiple variables applied with the new detection model and their occurrence importance that outperforms the existing solution.
This project shows an interesting case study in which Machine Learning applied to Big Data can have a noticeable impact on a company's turnover and therefore its profitability.
The graph shows the multiple variables applied with the new detection model and their occurrence importance that outperforms the existing solution.
Numbers speak for themselves. The solution led to a 25% increase in fraud detection. Furthermore, the False Positives to True Positives ratio saw a 33% improvement.
Our work
See more cases
Contact
Get in touch
Connect
