Grant: Discovery Development Grant
University of Guelph
$10,000.00
Robust Tools for High Risk Decision Exploration using Biomedical Data
FY 2017-18
Summary
A common analytical task is to make an involved (high-risk, or deeply impacting) decision based on a complex data domain. An important subset of these decisions can be made more informed by basing the decision on data from multiple samples within this domain. Tools to facilitate this type of analysis are not common, but have the potential to be widely applicable. I specifically see a need for clear visual tools to provide a decision maker a deep understanding of multi-sample data. This sort of data is common in medical studies where multiple samples derived from the same patient are integrated into a higher-level understanding; if this understanding can be structured to integrate well with other sources of data, a very valuable decision management strategy will be produced, and high-risk diagnostic decisions can be made more accurately, and more confidently. I am particularly interested in how best to provide this type of aggregate information to a decision maker. My primary area of application is in electromyographic data characterization for muscular disease state, but I am finding interesting applications of these techniques in other decision domains, particularly in postural data studies. I therefore propose to create a set of tools to aid in complex decision making. These will be based on a combination of machine-learning based analysis in the form of associative rule construction, coupled with a set of visualization tools to assist in deep understanding of the data. My industrial experience in visualization design will support the latter, and my several years of working with association mining (knowledge based rule discovery) guided EMG data analysis will support the former. These techniques can be validated in a variety of fields. I have a great deal of electromyographic disease data at my disposal, so this will provide an important area of application. Additionally this data can be applied to a variety of problems where time-series or other multiple-sample data is available.
Program
Discovery Development Grant
Recipient
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
DDG-2015-00007
Name nom
HamiltonWright, Andrew
Department département
Mathematics and Computer Science - Mathematics and Computer Science
Organization id
24
Institution établissement
University of Guelph
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2017
Competition year année de concours
2015
Award amount
10000
Program id
DDG
Program name en
Discovery Development Grant
Program name fr
Subventions à la découverte axées sur le développement
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
9999
Area of application group en
Not available
Area of application group fr
Non disponible
Area of application en
Not available
Area of application fr
Non disponible
Research subject code
9999
Research subject group en
Not available
Research subject group fr
Non disponible
Research subject en
Not available
Research subject fr
Non disponible
Partie
2
Nb partie
2
Application title
Robust Tools for High Risk Decision Exploration using Biomedical Data
Keyword
NSERC
Application summary
A common analytical task is to make an involved (high-risk, or deeply impacting) decision based on a complex data domain. An important subset of these decisions can be made more informed by basing the decision on data from multiple samples within this domain. Tools to facilitate this type of analysis are not common, but have the potential to be widely applicable. I specifically see a need for clear visual tools to provide a decision maker a deep understanding of multi-sample data. This sort of data is common in medical studies where multiple samples derived from the same patient are integrated into a higher-level understanding; if this understanding can be structured to integrate well with other sources of data, a very valuable decision management strategy will be produced, and high-risk diagnostic decisions can be made more accurately, and more confidently. I am particularly interested in how best to provide this type of aggregate information to a decision maker. My primary area of application is in electromyographic data characterization for muscular disease state, but I am finding interesting applications of these techniques in other decision domains, particularly in postural data studies. I therefore propose to create a set of tools to aid in complex decision making. These will be based on a combination of machine-learning based analysis in the form of associative rule construction, coupled with a set of visualization tools to assist in deep understanding of the data. My industrial experience in visualization design will support the latter, and my several years of working with association mining (knowledge based rule discovery) guided EMG data analysis will support the former. These techniques can be validated in a variety of fields. I have a great deal of electromyographic disease data at my disposal, so this will provide an important area of application. Additionally this data can be applied to a variety of problems where time-series or other multiple-sample data is available.