核心概念
Assessing predictability of fertility outcomes using Dutch survey and register data.
要約
The content discusses the importance of predicting fertility outcomes in the Netherlands using two datasets: the LISS panel survey data and Dutch register data. It introduces the PreFer data challenge aimed at improving understanding of fertility behavior through predictive modeling. The article outlines the methodology, phases, submission process, evaluation metrics, and criteria for determining winners.
Structure:
Introduction to Fertility Research
Explanatory vs. Predictive Modelling
Data Challenges in Scientific Progress
Benefits of Combining Survey and Register Data
Description of LISS Panel Survey Data
Description of Dutch Register Data (CBS)
Methodology of PreFer Data Challenge
Phases of the Challenge
Submission Process
Evaluation Metrics and Criteria for Winners
統計
LISSパネルは2007年から開始され、約5000世帯8000人の参加者を募集しました。
CBSデータには、1995年から2023年までの多くのデータセットが含まれています。
LISSデータとCBSデータを組み合わせることで、予測性能を向上させる機会が提供されます。
引用
"Out-of-sample predictive ability is an easy-to-understand measure of model quality."
"Data challenges have led to advancements in various scientific fields."
"Predictions based on survey data can be improved by augmenting it with register data."