Nina van Gerwen is a PhD candidate working at the Biostatistics department at the Erasmus Medical Center, where she will be researching how to derive causal predictions through a combination of machine learning techniques and joint models. She started her academic career with a Bachelor degree in Psychology, after which she focused her interests in statistics by pursuing a Research Master in Methodology & Statistics. After learning about the depth and breadth of statistics, she is incredibly motivated to continue learning more about it in different types of fields.

“Does ensemble learning lead to an improvement of prediction accuracy in settings with both longitudinal and time-to-event outcomes?”

In biomedical statistics, it often occurs that data consists of both time-to-event outcomes (e.g., time to death) and longitudinal outcomes (e.g., biomarker values measured annually), and then the interest lies in estimating a patient’s survival probability. There are currently diverse methods available to analyse this type of data. In this study I explore whether combining these different techniques through stacking, an ensemble learning technique from machine learning literature, can lead to an improvement in the prediction of survival probability.

PhD student,
Erasmus MC