Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Launch Supplement

University of Alberta

$12,500.00

Robust Statistical Modeling under Uncertainty

FY 2024-25

Summary
In recent years, our ability to collect, store and process vast amounts of data, together with rapid advances in technology, have led to the widespread adoption of data-driven decision-making. This includes the application areas, such as, shipment planning, where we predict the demand of product at warehouses using past product sales, weather forecasts or volume of Google searches for a product. In other areas, such as portfolio allocation, investor uses past security returns, analyst ratings, or behaviors of underlying securities to assign the ratio of portfolio to maximize the future return. More often than not, practitioners are forced to work with data obtained from various surveys. These problems, in statistics, can be formulized as conditional model by incorporating the side information or auxiliary variables. Most of existing methodology are built on the assumption that any future data all the future data we are going to encounter is distributed the same as our trained data. This assumption is not always valid. The first long term objective is to design a novel data-driven approach that is informed by observed side information to integrate side information, robust conditional estimation, and optimal decision under a distributionally robust framework. In particular, the new methodology will generate a more robust estimate with high efficiency and scalability in uncertain environment. We will also provide new theoretical understanding of the limitations of decision making problems in the face of multiple data sources. Our second long term objective is to find efficient computation method that is rich enough to capture the complexity of data and scalable enough to process massive datasets in a parallelized or fully decentralized fashion. The derived algorithms will be implemented in R to be available publicly. This proposal does fit into the NSERC portfolio. The novel methodology we proposed is critical, timely, and significant. This research will be a significant extension of the current scope of statistical estimation, both in terms of theory and applicability. It will also lay the foundation for the achievement of the long-term goal of building a bridge to conditionally statistical estimation problems and their inference, and enriching the computational toolbox and the theoretical developments for optimization under uncertainty.
Program
Discovery Launch Supplement
Location
Alberta, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
DGECR-2024-00318
Name nom
Wang, Yafei
Department département
Department of Mathematical and Statistical Sciences
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
12500
Program id
DGECR
Program name en
Discovery Launch Supplement
Program name fr
Supplément Tremplin vers la découverte
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1508
Committee name en
Mathematics and Statistics
Committee name fr
Mathématiques et statistique
Area of application code
1205
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Mathematical sciences
Area of application fr
Sciences mathématiques
Research subject code
3000
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Statistics and probability
Research subject fr
Statistique et probabilité
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Robust Statistical Modeling under Uncertainty
Keywords
distributional roubst; conditional estimation; data-driven decision-making
Application summary
In recent years, our ability to collect, store and process vast amounts of data, together with rapid advances in technology, have led to the widespread adoption of data-driven decision-making. This includes the application areas, such as, shipment planning, where we predict the demand of product at warehouses using past product sales, weather forecasts or volume of Google searches for a product. In other areas, such as portfolio allocation, investor uses past security returns, analyst ratings, or behaviors of underlying securities to assign the ratio of portfolio to maximize the future return. More often than not, practitioners are forced to work with data obtained from various surveys. These problems, in statistics, can be formulized as conditional model by incorporating the side information or auxiliary variables. Most of existing methodology are built on the assumption that any future data all the future data we are going to encounter is distributed the same as our trained data. This assumption is not always valid. The first long term objective is to design a novel data-driven approach that is informed by observed side information to integrate side information, robust conditional estimation, and optimal decision under a distributionally robust framework. In particular, the new methodology will generate a more robust estimate with high efficiency and scalability in uncertain environment. We will also provide new theoretical understanding of the limitations of decision making problems in the face of multiple data sources. Our second long term objective is to find efficient computation method that is rich enough to capture the complexity of data and scalable enough to process massive datasets in a parallelized or fully decentralized fashion. The derived algorithms will be implemented in R to be available publicly. This proposal does fit into the NSERC portfolio. The novel methodology we proposed is critical, timely, and significant. This research will be a significant extension of the current scope of statistical estimation, both in terms of theory and applicability. It will also lay the foundation for the achievement of the long-term goal of building a bridge to conditionally statistical estimation problems and their inference, and enriching the computational toolbox and the theoretical developments for optimization under uncertainty.