383. Mapping Financial Well-Being: A Bayesian Spatial Clustering Analysis of Regional Economic Disparities
Invited abstract in session MB-2: Optimization and applications, stream Nonsmooth and nonconvex optimization.
Monday, 10:30-12:30Room: B100/7011
Authors (first author is the speaker)
| 1. | Lok-Yung Wong
|
| Management Science, University of Edinburgh Business School |
Abstract
Conventional methods of measuring financial well-being, such as composite indexes and econometric models, lack the adaptability to reflect regional and behavioural variability. This paper presents a Bayesian-augmented DBSCAN framework to quantify clustering uncertainty by combining Monte Carlo Dropout with Bayesian Optimisation for adaptive parameter modification. Applied to English financial well-being data, the approach produces latent clusters representing geographic and behavioural aspects of financial resilience, savings, and borrowing stress. Unlike segmentation depending on stationary demographic characteristics, this method assesses confidence in every assignment by adjusting to local data structures. The main contribution is in creating a clustering approach that combines uncertainty estimation with density-based segmentation, thereby enabling more accurate and understandable detection of financial activity patterns. This paradigm facilitates transferable applications in policy targeting and behavioural risk assessment and is intended to increase the methodological rigidity of financial well-being analysis.
Keywords
- AI based optimization methods
- Optimization under uncertainty
- Optimization for learning and data analysis
Status: accepted
Back to the list of papers