Coronavirus (COVID-19) Updates and Resources »
Back to Top
Close Menu
 
Home » News-Events » News
 
News and Events Centre

UWI Research on Banking Sector

For Release Upon Receipt - Wednesday, May 5, 2021

A group of researchers from The University of the West Indies (UWI) has put forward solutions to specific problems in the banking sector that could improve customer experience and result in cost savings to financial institutions.

The grant project was funded by CIBC FirstCaribbean International Bank, with the research component undertaken by the St. Augustine Campus.

The Cave Hill Campus administered the grant under the guidance of Deputy Principal Professor Winston Moore, and supported by staff in the Office of Business Development – business development officer Sonia Johnson, projects assistant Rosita Spooner and assistant Julia Bailey-Smith.

The results were presented at an online forum on April 15, entitled Machine Learning Solutions to Problems in the Banking Sector.

Working closely with a bank, the team investigated issues such as marketing efficiency, chatbot assistance, fraud detection, digital channel recommender, and personalised overdraft protection. The collaboration with the financial institution ensured that the problem was accurately captured and the solutions were practical and useful.

Professor of Computer Science Patrick Hosein supervised the work of lead data science researcher Shiva Ramoudith, lead researcher Reshawn Ramjattan, PhD candidate in computer science Shellyann Sooklal, and MSc students in data science Akeel Mohammed and Karesia Ramlal.

“The benefit to the customer is more along the line of better service. All these projects work in the back end for the banks, so the customer doesn’t necessarily see the inner workings, they just get the results of it,” Mohammed explained.

He focused on the topic ‘Optimising Digital Channel Profitability’, noting that given the traditional in-person approach to banking in the Caribbean, online banking platforms would not be fully developed if there was an inadequate number of customers utilising such services.

He specifically examined how digital channels could benefit from optimisation to improve adoption and, thereby, profitability.

Sooklal presented A Credit Card Fraud Detection Algorithm to Improve the Accuracy and Speed of Detection, a project that seeks to improve the accuracy of real-time online and offline detection systems.

“What we tried to do is improve the accuracy of being able to tell if a transaction was fraudulent. So if it was not them [the customer] who made the transaction, we want to be able to detect that so they do not run into problems later on when they realise there was a fraudulent transaction on their account,” she said.

Lead data science researcher Shiva Ramoudith investigated how banks can make marketing campaigns more effective.

He cited inherent problems in the methods currently used, such as the need to contact some customers multiple times for them to agree to a service while others may become irritated if too many contacts were made and could leave the company as a result.

Ramjattan presented Conversational Interfaces of Banking Services, known as chatbots.

The problems he highlighted revolved around learning to use such apps and websites for banking services, with numerous apps competing for user attention and resources.

He noted that apps need to show their value immediately and on a consistent basis to be utilised to their full extent.

Ramlal presented on Personalized Overdraft Protection Framework, notably how clients could better benefit from overdraft protection, as she pointed out that market research suggested that existing options do not meet the needs of clients.

She stated that customers have identified the existing cost structure and authorised limits as the main reasons for not using the service.

At the end of the presentations, Professor Hosein made a case for the region to produce more data scientists.

Hosein, a leading authority on data science, has received numerous accolades, including the Huawei US Wireless Research Employee of the Year in 2007 and being named a 2015 Anthony Sabga Caribbean Laureate.

He obtained five degrees at the Massachusetts Institute of Technology, one of which was his PhD in electrical engineering and computer science.

“It’s so very important that the region develops expertise in data science, which is more widely known as artificial intelligence (AI) but is actually a subset. Data science can be applied to a wide range of problems in [many] industries such as insurance, banking, telecoms, et cetera. It can be used for decision-making in government, to improve education, and the list goes on.”

Professor Hosein has also published extensively and holds 40 patents in telecommunications and wireless technology. He believes the region has the talent to solve its own problems.

“This means that in the coming years there’ll be a huge demand for high quality data scientists, so we need to start producing such scientists. The solutions we need for our regional AI-related problems would typically be unique because of our unique circumstances, our unique culture, society et cetera and we cannot simply buy off the shelf solutions.

“Secondly, the data sets required for machine learning solutions algorithms must reflect our society, our environment, our small size, and hence, we need to collect these ourselves and not rely on data sets collected from developed countries.”










Publications

Cave Hill Campus Chill Magazine
Browse our collection of publications, covering life at all four of our UWI campuses - Cave Hill, St. Augustine, Mona and Open.

Cave Hill


Reach Us

Have more questions?

Talk to The Office of Marketing and Communications
The Leslie Robinson Building

 

Email: publicinformation@cavehill.uwi.edu
Tel: (246) 417-4000
Fax: (246) 417-4000
Website: www.cavehill.uwi.edu/news-events

 

Office Hours:
Monday to Friday
8:30 a.m. - 4:30 p.m