Machine Learning in Credit Risk: Performance, Interpretability, Statistical Significance

Research seminar

Date: September 13

The seminar will start with Machine Learning basic formulations. Then we will explore an example predicting the Probability of Default based on an openly available dataset. High-stakes models have to be transparent for auditors, but the performant ML models are inherently opaque. Thus, we will describe methods for model interpretation and confidence intervals for some of the interpretation results.

Speaker: Kyrylo Medianovskyi, Data Scientist at Group Model Risk and Validation. Swedbank AB Estonia.

The seminar will be conducted in English.

September 13 d. 15:00

Gediminas st. 50 Conference hall

Įtraukti į iCal
Suggest an Event