1. Research & Development Dr. Kamran Sartipi PhD, P.Eng
RESEARCH & INNOVATION
Kamran Sartipi, Ph.D., P.Eng.
My research projects have been inspired from the vision I acquired through several years of
collaboration with researchers in Engineering, Business, Heath Sciences, and Entrepreneurship. I have
served as a program committee member for 34 international conferences, and external reviewer for 15
journals. I am a co-investigator of an Ontario Research Fund (ORF) project with total funding of $1.4
Million and I have secured prestigious NSERC Discovery Grants until 2019. I am serving on the Peer-
Review Committee for the CHRP (Collaborative Health Research Projects), which is a granting
organization with collaboration of CIHR (Canadian Institutes of Health Research) and NSERC.
I have contributed to the complementary inter-disciplinary research fields, listed below:
• Software Engineering & Data Analytics (34 publications)
o Dynamic and Static Analyses of large Software Systems, User Behaviour-Pattern Discovery
from Large Datasets, Data Mining, Service Intelligence, and Semantic Analysis.
• Information Security & Healthcare Informatics 30 publications)
o Security and Privacy in Distributed Systems, Clinical Decision Support Systems, Cloud and
Mobile Service Interoperability.
My research projects incorporate both theoretical and practical aspects, which are supported by working
prototype tools. As a part of my PhD research, I have developed a graph-pattern-matching based
software architecture recovery technique and a toolkit namely “Alborz” with 35,000 lines of code. I
have been a pioneer researcher in applying data mining algorithms on both static and dynamic analyses
of software systems. In the followings, first the snapshots of my research projects are presented, and
then two large projects that I currently pursue will be introduced.
Domain-independent & Domain-specific Research Projects
Overall, my research projects are categorized as “domain-independent” and “domain-specific” (Medical
Systems and Health Informatics). Below, the scientific aspects of these projects are summarized. In each
project, references to the publications in my CV are indicated, where: b:book; ch:chapter; j:journal;
js:journal-submitted; and c:conference.
Domain-independent research projects
• Enhancing service oriented architecture (SOA). This research enhances data privacy/security,
reduces network traffic, and provides new enterprise level features. It introduces two new concepts
“task service” and “ service representative” in the SOA environment. Task service is a multi-
component (model, knowledge, data) web service that can process the client data locally at the client
side. Service representative is a generic agent at the client side that will be customized by the
knowledge component and will execute the model component on both client and task service data.
This approach will enhance the SOA architecture in different ways [j6, js2, js1, c39, c38, c36, c35,
c34, c32].
• Information Security and privacy provisioning. This research (mentioned above) proposes a SOA-
based architecture that provides centralized security among DI-r, distributed PACS and mobile
clients in cloud environment. Key challenges with this trust model are: lack of federated capabilities
to ensure unique user identity in legacy domain; preserving consistency of access control rules
Page ! of !1 4

2. Research & Development Dr. Kamran Sartipi PhD, P.Eng
across all systems, and monitoring user-behavior patterns to enhance security policies. [j10, j8, js4,
c48, c47, c46, c45, c44, c42, c22].
• Software system analysis and architecture recovery using data mining. This research provides in
depth knowledge about legacy software systems and their architectures. In this respect I have done
extensive research in both dynamic and static software system analysis by applying data mining and
pattern matching techniques.
o Dynamic Analysis. This research identifies the implementation of specific software
functionality within a software system without any prior knowledge about the source code.
The approach consists of applying specific sets of scenarios on an instrumented software
system to extract execution traces. Next, sequential pattern mining algorithm and concept
lattice analysis are applied to extract execution patterns and locate the target source code. We
expanded this approach by applying it on service-oriented architecture (SOA) to measure the
quality of web services in service selection and composition. [j5, c37, c26, c23, c20, c18, c17,
c16, c15, c14, c12, c27].
o Static Analysis. This research addresses the design and development of an incremental
software architecture recovery and evaluation environment using data mining techniques.
The environment is interactive and provides: pattern-based architectural recovery using a
query language and approximate graph pattern matching; optimization clustering;
partitioning; and view-based architectural design evaluation. These techniques have been
implemented within my Alborz toolkit, which are mentioned below [b1, ch1, j3, j1, c24, c10,
c9, c8, c7, c6, c5, c4, c3, c2, c1].
Domain-specific research projects (Medical Systems & Health Informatics)
Research in eHealth and medical infrastructure encompasses different areas such as software
engineering, information technology, web services, security and privacy, and medical and healthcare
standards. In particular, my research team has contributed in the following areas:
• Smart decision support systems. This research aims at providing a new generation of decision
support systems where mined-knowledge at decision points (as reminders, alerts, recommendations)
will assist the physicians (for patient diagnosis) or administrators (for resource allocation) to make
effective decisions. In this context, mined-knowledge refers to the extracted patterns and trends
from clinical data using data mining techniques. This research covers both rule-based and flow-
based decision support techniques [j9, j4, ch3, ch2, c43, c41, c31, c13, c11].
• Knowledge translation to HL7 v3 messages. This research provides a novel and well-defined
process to guide translation of healthcare transactions (obtained from healthcare scenarios) to HL7
v3 Interactions. The approach re-categorizes HL7 v3 Interactions based on their behavioral traits in a
messaging context and uses semantic analysis to automate the process of identifying HL7 v3
Interactions. This approach has been implemented in the TAMMP tool described below [j7, j2, c40,
c30, c25, c21, c19].
• Mobile eHealth. This research provides efficient techniques to integrate different devices such as cell
phones, tablets, and specialized devices such as pacemakers to be used seamlessly with other
software services of the electronic health record (EHR) systems. In this approach data will be
collected, maintained, analyzed and communicated using HL7 information models and messaging.
[c35, c34].
• Cross-domain interoperability. Most service interoperability techniques focus on standardization and
interoperability among information systems within the same domain. In this research we address the
challenges in exchanging information among heterogeneous applications in different domains, such
Page ! of !2 4

3. Research & Development Dr. Kamran Sartipi PhD, P.Eng
as healthcare and insurance. The approach is based on a domain-independent reference information
model as a basis for information exchange at the syntactic and semantic levels. [js3, c29, c28].
I have brought together the above research activities into two large projects that are covered within a
Multi-disciplinary Research Centre for Medical Systems and Informatics that I initiated at UOIT:
Project 1. User Behavior-Pattern Discovery and Intelligent Security Provisioning in Distributed
Systems
This research is underway as a part of the ORF (Ontario Research Fund) project with collaboration of
McMaster Department of Radiology, Mohawk College MEDIC lab, and NexJ Health. The architecture
of this research project is shown below. My research team is working on the following projects:
• Intelligent middleware security provisioning. A secure, central and service-based “intelligent
middleware” consisting of: multi-agent technology (smart local agents and administrative
middleware agents) for two-level decision-making process; central policy repository and
management; central metadata repository for images; and a centralized authentication and
decentralized authorization model.
• Knowledge-driven user behavior-pattern discovery. This research provides new generation of
intelligent decision support systems that effectively assist the system administrators to obtain
deep insight into the system user's dynamic behavior patterns in order to refine the existing
security policies using a novel behavior pattern query language (BPQL).
• User behavior Simulation environment. An event-log generator engine is developed which
receives administrator-defined user-behavior patterns using a pattern language and produces
corresponding events in the context of noise events which allows us to effectively test and fine-
tune the above techniques before applying them on the production event-logs. Currently we are
experimenting with the audit logs from a distributed medical imaging system running at Mohawk
College.
Project 2: Secure Multi-Datacenter Network for Data and Service Interoperability, Knowledge
Extraction and Dissemination among Medical Specialty Domains.
This project is underway with collaboration of the Department of Medicine (Gastroenterologists) at
McMaster University. The architecture of the multi-datacenter is shown below.
In this project, we develop a data center for the McMaster IBD (inflammatory bowel disease) Clinic to
extract knowledge from patient data that will be scaled up and applied to a multi-center collaboration to
support a network of ambulatory care IBD clinics. Each datacenter is autonomous in terms of the
adopted EHR (Electronic Health Record) system, data schema and communication protocols, and will
be implemented using a private cloud infrastructure. The integration among the centers are based on
standards, such as SOA (SOAP or RESTfull services), HL7 v3 (medical messaging standards)
SNOMED CT (concept interoperability), and PMML (knowledge interoperability). The system security
and intelligence will be provided through OpenID Connect authentication and authorization technology.
Tool Development
My research team has contributed to the practical applications of knowledge, through design,
development, documentation, and presentation of several software toolkits, as follows:
▪ Consultant-as-a-Service: A Context-driven Mobile Decision Support Service (2011-2013).
Page ! of !3 4

4. Research & Development Dr. Kamran Sartipi PhD, P.Eng
http://faculty.uoit.ca/sartipi/Posters/2013CASCONPoster.pdf
Consultant-as-a-Service (CaaS) is a new generation of decision support services using techniques
from software agents, semantic analysis, and data mining. CasS is offered as a set of mobile and
cloud APIs that coach the user, who is not familiar with an organization, to effectively select desired
organization's business services.
▪ TAMMP: Tool-Assisted Knowledge to HL7 v3 Message Translation (2007-2009).
http://faculty.uoit.ca/sartipi/eHealth/TAMMP/Home.html
The TAMMP tool guides system developers to design the required communication messages within
different healthcare integration projects. It allows the developers to translate healthcare scenarios
into HL7 v3 messages in a seamless manner. The mapping process is based on the concepts
“structured healthcare transactions” and “contexts”. TAMMP uses a search tool built on leading edge
Semantic Web technologies to automate mapping of structured healthcare transactions onto HL7 v3
Interactions.
▪ MacSeie: McMaster Service-based eHeatlh Integration Environment (2009-2010).
http://faculty.uoit.ca/sartipi/eHealth/MacSeie/Home.html
The MacSeie environment allows interoperability between two legacy healthcare systems using
service oriented architecture (SOA) and HL7 v3 messages. The first system is a clinical decision
support system (known as Vascular Tracker, VT) from the COMPETE III research team at St. Joseph
's Healthcare Research Center in Hamilton. The second system is an electronic medical record
system (EMR), which contains the medical information of patients. The VT and EMR systems are
web-based and accessible through a web browser.
▪ ALBORZ: Software Architecture Recovery using Data Mining (1998-2007).
http://faculty.uoit.ca/sartipi/Alborz/static/index.htm
(1998-2003) I developed Alborz toolkit as part of my PhD research to perform pattern-based
software architecture recovery and evaluation. The original version was implemented in “Refine”
reengineering environment (+35,000 lines). During 2004 to 2007, I migrated the original tool from
“Refine” environment into Eclipse IDE platform using plug-in technology and enhanced the user-
interaction of the tool. I have demonstrated both static and dynamic techniques of the Alborz toolkit
in several international conferences, workshops, and technology showcases.
Page ! of !4 4