Ordine Attuari

Leggi il contenuto della pagina | Accesso Rapido | Contatti



Albo, esame, oggetto della professione

Area Riservata

Accedi all'area riservata agli iscritti

Offerte di lavoro

Solo per iscritti, segui le offerte

WEBINAR INVITATION: Credit Risk Modelling-26 September 2019Feed RSS

Corsi di Formazione - 12/09/2019

On behalf of the International Actuarial Association's (IAA) Banking Working Group and the Actuarial Society of South Africa's (ASSA) Banking Committee, you are invited to participate in the upcoming webinar entitledAutomatic Relevance Determination Bayesian Neural Networks for Credit Card Default Modelling.The webinar will be held on Thursday, 26 September 2019 at 10:00 EDT // 14:00 GMT // 16:00 SAST.The duration of this event will be for a total of 90 minutes with a 30-minute question and answer session.


The keynote speaker will be Rendani Mbuvha.


Rendani is a fellow of the Actuarial Society of South Africa and a holder of the Chartered Enterprise Risk Actuary designation. He holds a Bsc Honours in Actuarial Science from the University of Cape Town and an Msc in Machine Learning from KTH, Royal Institute of Technology in Sweden. Rendani currently works as a lecturer in the School of Actuarial Science and Statistics at the University of Witwatersrand while pursuing his PhD in Artificial Intelligence at the University of Johannesburg. Rendani is a recipient of the 2019 Google Africa PhD fellowship which supports his PhD work in Bayesian Neural Networks. He has previously worked as an Actuary at Discovery, Milliman and Barclays Africa Group.



Credit risk modelling is an integral part of the global financial system. While there has been great attention paid to neural network models for credit default prediction, such models often lack the required interpretation mechanisms and measures of the uncertainty around their predictions.

This work develops and compares Bayesian Neural Networks (BNNs) for credit card default modelling. This includes a BNNs trained by Gaussian approximation and the first implementation of BNNs trained by Hybrid Monte Carlo (HMC) in credit risk modelling. The results on the Taiwan Credit Dataset show that BNNs with Automatic Relevance Determination (ARD) outperform normal BNNs without ARD. The results also show that BNNs trained by Gaussian approximation display similar predictive performance to those trained by the HMC. 

The results further show that BNN with ARD can be used to draw inferences about the relative importance of different features thus critically aiding decision makers in explaining model output to consumers. The robustness of this result is reinforced by high levels of congruence between the features identified as important using the two different approaches for training BNNs. 

Participant Registration:


To register for this event, please click on the following link:

 ·           Click here to be automatically forwarded to the webex registration 

Once successfully registered, you will receive an automated response which will include a link to use to login on the day of the webinar.


On the day of the event, you will need a stable internet connection to a computer device. For the audio portion of the presentation, a telephone will not be required however, speakers, a headset or ear buds plugged into your computer device will be needed. Questions during the Q&A portion of the presentation can be typed into the application on your computer device.


IMPORTANT NOTE:It is important for participants to know whether they will have firewall restrictions (via employer facilities) for the computer they plan to be using to participate, as you may very well experience difficulties in connecting or logging into the presentation.


*This event may be of interest to those working in the Banking industries. You are welcome to forward this invitation to anyone who you know that may find this webinar to be of interest and would like to participate.


Should you have any questions, experience any difficulties or need anything further, please send an email directly to technical.activities@actuaries.org.


Thank you for your attention and participation.