Prof Tanja Verster‘My area of interest and expertise, both in teaching and research is predictive modelling – specifically industry-focused teaching and industry-focused research. Predictive modelling is a mathematical process used to predict future events or outcomes by analysing patterns in each set of input data. Many predictive models are used in the financial industry, ranging from regulatory models for credit risk (Basel and IFRS 9) to fraud detection models, customer retention models, collection strategies models, and market risk models. Yet these predictive models are not only restricted to the financial industry – the areas applicable are limited only to your imagination.’

The range of the applications of the models mentioned by Professor Tanja Verster of the Centre for Business Mathematics and Informatics (BMI) at North-West University (NWU) is very large: some examples include ‘predictive modelling based on the subscriber’s viewing preferences (for streaming services), healthcare diagnostics, weather prediction, fraud detection and sales forecasting.’

She continues: ‘BMI spans multiple disciplines, integrating Mathematical Science, Information Technology, and Economics. Rather than fitting into a single discipline, it encourages interdisciplinary collaboration.’ She remarks that NITheCS fosters this integration and collaboration between researchers from different fields and institutions. ‘A collaborative structure highlights the interdisciplinary nature and relevance of BMI’s contributions and aligns BMI’s research with the Quantitative Finance and Data Science domains. Because NITheCS provides multiple opportunities for researchers to participate in research programmes and attend seminars, colloquiums and summer schools, I deeply appreciate the interdisciplinary environment within which NITheCS operates.’

Background and influences
Tanja launched her career with a Master’s degree in Quantitative Risk Management – initially as a quantitative analyst at First National Bank, before transitioning to academia as a lecturer at NWU in May 2003. She then earned her PhD in Risk Analysis with a focus on credit scoring from the Centre for BMI in 2007. Currently, she teaches postgraduate courses in credit scoring, predictive modelling and data mining, and engages in applied research projects related mostly to predictive modelling in credit. ‘Over the past 22 years at BMI, I have published over 30 peer-reviewed papers, completed more than 40 industry-directed research projects, and supervised numerous PhD and Master’s students. I was recognised as a C-rated researcher by the National Research Foundation (NRF) in 2020.’

She singles out two people as highly influential in her career choices and progression: ‘My mom passed away just after I handed in my Master’s degree – thus missing my graduation, which was particularly sad for me. As a single mother of four, she had tirelessly supported us all. She was my cheerleader – always proud of my accomplishments. My gratitude also goes to Prof Riaan de Jongh as my mentor for two decades. His unwavering focus on industry-focused research and teaching profoundly influenced my own academic journey.’ A deeply religious Christian thanks to her mother’s influence, Tanja adds that she is ‘incredibly proud to see my 17-year-old twin daughters embracing their faith as they navigate their own journeys.’

Practical examples and machine learning
Examples of their research implemented at banks include the paper ‘A Proposed Benchmark Model Using a Modularised Approach to Calculate IFRS9 Expected Credit Loss’ (2020). This industry-focused project paper outlines a transparent, modularised approach to calculate Expected Credit Loss (ECL) under IFRS 9, using components like probability of default, loss given default, and exposure at default. It provides a benchmarking tool for IFRS 9 model development. Another IFRS 9-related paper, ‘A Forward-Looking IFRS 9 Methodology, Focussing on the Incorporation of Macroeconomic and Macroprudential Information into Expected Credit Loss Calculation’ (2023), proposes a method based on principal component regression to adjust probability of default (PD) term structures. Both of these methodologies have been implemented within the banking sector.

Among other examples is the paper ‘Developing an Impairment Loss Given Default Model Using Weighted Logistic Regression Illustrated on a Secured Retail Bank Portfolio’ (2019). It introduces a novel approach to modelling Loss Given Default (LGD) under IFRS 9 and proposes a two-component methodology for estimating LGD in non-default and default accounts. ‘The methodology, which has been implemented within a South African bank, strengthens predictive modelling applications in credit risk,’ explains Tanja. Related works, such as ‘A Motivation for Banks in Emerging Economies to Adapt Agency Ratings When Assessing Corporate Credit’ (2019) and ‘Development of an Impairment Point-in-Time Probability of Default Model for Revolving Retail Credit Products: South African Case Study’ (2021), continue this practical focus, with the latter also implemented by a local bank.

Tanja says the landscape of financial credit risk models is changing rapidly. ‘Many factors influence financial credit risk modelling. For example, let’s consider machine learning. As machine learning expands, it becomes necessary to understand how these techniques work and how they can be applied. Another factor is financial crises. Where predictive models view the future as a reflection of the past, financial crises can violate this assumption. This creates a new field of research on how to adjust predictive models to incorporate forward-looking conditions, which include future expected financial crises. Furthermore, consider the impact of financial technology (Fintech) on the future of predictive modelling. Fintech creates new applications for predictive modelling and therefore broadens the possibilities in the financial predictive modelling field. This changing landscape causes some challenges but also creates a wealth of opportunities. Indeed, there is a lot of overlap between machine learning and statistics. Most machine-learning techniques used in the predictive modelling context relates directly to statistics.’

Tanja quips: ‘Machine learning can be equated to a gun: in the wrong hands, it is terrifying, but in the right hands, it is a powerful tool. When machine learning is used in predictive modelling, it should be combined with a statistical basis and business knowledge. This will ensure that machine learning becomes a powerful tool.’