Mastat Day 2020

For whom
Alumni
When
19-02-2020 from 12:00 to 18:00
Where
De Krook, Miriam Makebaplein 1, 9000 Gent, room De Blauwe Vogel (level -1)
Language
English
Organizer
Jan De Neve
Contact
Jan.DeNeve@ugent.be
Add to my calendar

The Mastat Day brings together students, alumni, lecturers and stakeholders.

Program:

12:00-13:00 Pre-defense poster session with sandwiches
13:00-14:00 “Designing and implementing a survey: Theory and practical issues” dr. Joost Kappelhof
14:00-15:00 “Machine Learning and Causal Inference: Tools for Data Science” by dr. Karla Diaz-Ordaz
15:00-15:30 Coffee break
15:30-17:00 Alumni presentations

  1. Katrien Baert
  2. Tom Everaert
  3. Barbara Cornillie
  4. Joris Meys

17:00 Reception with pre-defense poster session

Please confirm your attendance before Februari 11th

 

Abstracts

Designing and implementing a survey: Theory and practical issues” dr. Joost Kappelhof

Using survey data to draw inferences about a population under study is a challenge. Many different error sources can potentially affect the data quality. This can in turn lead to wrong conclusions about the population under the study. In this presentation I will present a broad overview of the challenges and trade-offs that have to be made when designing and implementing a survey to increase awareness among data users about the data quality.

 

Machine Learning and Causal Inference: Tools for Data Science” by dr. Karla Diaz-Ordaz

Health data scientists and statisticians are often consulted by medical and policy decision makers evaluating the effects of medicines or social programmes (henceforth referred to as treatments), who need reliable evidence about the long-term causal effects of competing treatment strategies in key populations. Machine learning (ML) methods have received a lot of attention in recent years within the data science community, especially in settings with a large number of variables. However, causal effect estimation often involves counterfactuals, and prediction tools from the ML literature cannot be readily used for causal inference. In the last decade, major innovations have taken place incorporating supervised ML tools into estimators for causal parameters such as the average treatment effect (ATE). This holds the promise of attenuating model misspecification issues, and enhancing researchers knowledge with variable selection.
In this talk, I will review some of these developments incorporating machine learning in the estimation of the ATE of a binary treatment, under the “no unobserved confounding” and positivity assumptions. In particular, I will illustrate (1) an ensemble machine learning algorithm known as Super Learner, for prediction and (2) estimation of ATE with machine learning approaches, such as targeted maximum likelihood estimation.
Throughout, I will use as illustrative example the evaluation of cancer immunotherapy in Non-Small-Cell Lung Cancer patients, using electronic health records and tumour genomic data from a large USA cohort.

 

Find more events: