Address:
Department of Computing
and Software
(Rm. ITB/223)
McMaster University
1280 Main St. W.
Hamilton, Ontario
L8S 4K1, Canada
Tel (office):
905-525-9140 Ext. 23101
E-mail:
Welcome to my
homepage,
On this page, you can find relevant information about my
research activities prior to August 2006. If you have questions, please feel free to
contact me.
In my current research, I
am working on interoperable provision of clinical mined knowledge
at the point of care. The focus is on the clinical knowledge that is obtained
through some sort of Data Mining activities, e.g., diagnosis
of a disease using a decision tree classifier; clustering of
patients based on their risk factors; or gathering new
knowledge in the form of association rules from volumes of
past patients' cases and history. Interoperable, on
the other hand, is a very common Software Engineering
buzzword. In this context, it means the situation in which
adaptation of the knowledge that is produced elsewhere,
requires minimum effort at the Healthcare institutions that
are eventually going to use it. My Master of Applied Science
thesis on this topic is available
here.
This research requires a
true multi-disciplinary approach. The clinical knowledge is
encoded using the PMML specification and used to guide the
practitioner at decision points in the guideline flow. The
clinical guideline model that is based on the
Guideline Interchange Format3
(GLIF3) specification is extended to include
data-mining-model-based decision making.
In the
publication section, you can find more on this topic.
Also in the tools and development section, you can have
access to the Guideline Execution Environment (GEE) binaries
and source code.
I have studied the
application layer multicasting for multimedia
streams. My research involved proposing a congestion
control mechanism that takes advantage of special
video encoding techniques (transcoding) to improve
QoS measures. I implemented it in the Network
Simulator-2 (ns-2)
protocol stack and carried out performance and QoS
simulation analyses. This work formed the basis of
my Bachelor of Science thesis in Computer
Engineering that I received from
Sharif
University of Technology
in 2004.
I have also had a
taste of mobile/embedded systems research and
development experience. This involved adopting
smartcards (Java
cards)
for patient identification and storage of his
medical records, along with Java card-enabled
handheld PCs that provide mobility for their
caregivers.
R. Sherafat, "Interoperability of Data and Mined Knowledge in
Clinical Decision Support Systems", Master
of Applied Science in Software Engineering thesis,
Department of
Computing and Software,
McMaster University, August 2006
GEE (Guideline
Execution Environment):
Guideline Execution
Environment (GEE) is developed as an extending plug-in to
the widely used Protégé ontology editor tool. The
Guideline Interchange Format3
(GLIF3) is
the base specification that has been extended to incorporate
new data mining-based decision support features and
constructs. Apart from GEE which is the execution platform on Protégé, a
slot widget is also developed to handle guideline modeling
accommodating the extended GLIF3 data mining features.
These two plug-ins are written in Java and use the standard
Protégé Ontology API to access the guidelines knowledge base.
Simulation of an
overlay networks congestion control mechanism for
multicasting video
streams in ns-2:
The flexibility of overlay
networks and the application layer .. makes them an ideal
approach to handle complexities involved with the QoS
requirements of video streams. In this simulation, I have
studied and analyzed a congestion control approach with
applications to video multicasting that took advantage of a
special video encoding technique (transcoding) to improve
QoS.
July 2006,
Interoperability of Data and Mined Knowledge in
Clinical Decision Support Systems, Master of
Applied Science defense presentation, McMaster
University (ppt)
June 2006,
Incorporating Data Mining Applications into Clinical
Guidelines, Computer-Based Medical Systems,
19th IEEE Computer-Based Medical Systems 2006, Salt
Lake City, Utah (ppt)
May 2006,
Web Servers: The Engines that Drive the World Wide
Web, Accelerated Students Workshop, McMaster
University (ppt)
October 2005, Interoperability & Knowledge Management in
Healthcare, Communications and Information
Technology Ontario, Ottawa (Presented by Dr. Kamran Sartipi
-
ppt)
February 2005, Requirements Docuementation, CAS 703
course seminar, McMaster University (ppt)
November 2004, Knuth-Bendix Completion Algorithm, CAS
701 course seminar, McMaster University (ppt)