Professor Keming Yu
Professor

Research area(s)

1. Building on mathematical theory and data-analytic methodologies, my research develops statistical models, methods, and optimal algorithms to address key challenges in the following areas:

   • Novel regression models and methods, including quantile regression for risk analysis in financial econometrics and business applications.

   • Robust algorithms for machine learning and deep learning, with an emphasis on stability, interpretability, and reliability.

   • Statistical analysis and machine learning for modeling loneliness and social isolation in gerontology, integrating methodological innovation with real-world health and social data.

   • Extreme regression and random forest methods for analyzing inequality and evaluating interventions, including inequalities in health, public services, income, education, wellbeing, obesity, and biomedical outcomes.

   • Statistical and machine learning approaches for engineering risk assessment, such as compound weather and climate events, flooding, cable faults, pipeline corrosion, and wind turbine reliability.

   • Statistical theory and methodology, including Bayesian analysis, for both big data and small data settings.

Grants

Project details

Past Grants as PI:

The UK Office for National Statistics (ONS) awarded funds in its first ever open call for research in economic measurement to him for a project entitled New regression models for the analysis of wellbeing and income distribution (https://www.brunel.ac.uk/research/Projects/New-regression-models-for-the-analysis-of-well-being-and-income-distribution)  (04/2019---03/2020), £5000.00.

EPSRC Case Studentship: Data Analytics for Risk Based Decision Making in Asset Integrity Management (01/01/2018—30/09/2020), £80,670.

National Institute for Health Research (NIHR): NIHR Research Methods Opportunity Funding Scheme £70,517 (2010—2011) and £68,883 (2014-2015).

NSIRC scholarship funded by TWI (Cambridge) for supporting  three  PhD study on Statistical analysis and modelling of corrosion data:  £98,890 (2014—2020).