- Lecturer: Wang, Jinfang
*Lecturer Information: https://w-rdb.waseda.jp/html/100003624_ja.html - Title: Personalized Data Management Science
- Day and Time: January 18, 2024 6:30PM
- Abstract:
Data Science encompasses two distinct cultures: traditional model-based Statistics, known for its interpretability, and the more recent algorithm-driven Machine Learning (ML), renowned for its predictive prowess. Traditional Statistics primarily aims at deducing population traits from sample data, whereas ML concentrates on forecasting ‘typical’ future observations.
In this talk, I will present some recent thoughts on integrating these two cultures, focusing on individuals with specific adverse conditions, like pre-diabetes, to improve their status by modifying manageable factors, such as lifestyle habits. The proposed personal management program, for instance, aims at transitioning individuals from pre-diabetes to a normal state by iteratively utilizing statistical inference and ML prediction. Statistical inference identifies variables for management and ML helps for outcome prediction based on updated counterfactual characteristics.
This novel approach, termed ‘Personalized Data Management Science’ (PDMS) leverages data science for enhancing individual well-being. PDMS’s diverse potential applications include, but are not limited to, personalized health management, adaptive learning, tailored athletic training, customized interventions for preventing juvenile delinquency.
In this talk, I will focus on a detailed case study within the healthcare context, specifically on pre-diabetes management, to illustrate the effectiveness of this proposed framework.
