Avan, S.Si., M.Si., an alumnus of the Mathematics program and a Risk Modeler at Bank Mandiri, began his presentation by discussing the current state of the banking sector, especially in risk management. The talk was part of a guest lecture for the Artificial Neural Networks course with the theme “Credit Risk Modelling” held on December 10, 2024, at the Mathematics Department. With nearly a decade of experience as a risk modeler at Bank Mandiri, Avan shared best and worst practices for building models. Most of the attendees, who were students, paid close attention as they were familiar with the model-building framework, similar to their final thesis projects in the Mathematics Department.

“Next, we’ll discuss issues related to imbalanced datasets and overfitting,” Avan said. These two problems often challenge mathematics students when working on their theses. He explained these issues in a clear and detailed way. The lecture also covered artificial neural networks, which are a popular machine learning model used in risk modeling. The guest lecture was held in a hybrid format, with both students and lecturers from the Mathematics Department of FMIPA UNEJ attending.

During the discussion session, Avan answered questions from students, including Ade and Deka. They asked about how to handle imbalanced datasets effectively and how to build an effective neural network model. “Imbalanced datasets are common in banking, and there are several techniques and tips to overcome this issue and still produce a good machine learning model,” he explained. Regarding building a good machine learning model, Avan suggested adjusting the hyperparameters properly to create the best model.

Avan emphasized that each case requires a different approach, so understanding the specific case deeply is crucial. At the end of the session, he advised the students not to focus only on developing theoretical knowledge or skills. “It’s important to sharpen technical skills, such as coding and the ability to generate insights, as both are valuable and in demand in the job market,” he concluded.