|
|
|
|
People
Faculty member of Computer Engineering Department Sharif University of Technology.
|
 |
current members
- Thesis : "Data Stream Clustering"
|  |
- Thesis : "Search Results Clustering"
|  |
- Thesis :"approximation algorithms for ontology mapping extraction"
|  |
- Thesis :"A new model for Context-Aware Semantic Web Services"
|  |
former members
- Mehdi Hosseini (MS Student)
- Thesis Title: "Learning User Need For Improve Search Result In Web"
Nowadays, various services are provided for internet users by analyzing and extracting knowledge from search engines. For example, many search engines exploit knowledge extracted from query logs for improving search results. Categorizing queries based on concepts is the most significant step, so in this thesis we first introduce new approach for clustering queries and selected URLs concurrently, then two applications of it are demonstrated. In first application a query recommender system is introduced using clustered queries and related URLs. In the other application clustered queries and URLs related to a special web site are applied to analyzing the information architecture and the link structure of the web site.
|  |
Thesis: ": Ontology Learning and Taxonomy Construction from Text Corpus"
|  |
Thesis: "News Categorization and Ranking"
Nowadays, there are plenty of online websites related to news. Hence, new technologies, tools and special search engines are created for having access to the news on these websites. Online news is a special type of public information which has exclusive characteristics. These characteristics contribute news engines tasks such as discovering, collecting and searching to be different with similar tasks in traditional web search engines and also clustering plays conspicuous role in news engines tasks. We try to find new models for Ranking and Categorizing news engines.
|  |
Thesis: "Mining the Persian Blogs' Social Network"
Nowadays users of the Web are encouraged to generate content on the Web by themselves. Weblogs are the most convenient tools to perform content generation. In fact, Blogsphere is a kind of social networks and Weblogs are one of the most important components in Web 2.0. I this thesis, we analyze the link structure of the Blogsphere to discover the patterns and trends among Weblogs. Our analysis is based on different aspects of Data Mining algorithms. One of the most important applications of the hyperlink analysis of The Blogspace is inferring communities. Another meaningful issue is exploiting the special link structure of the Weblogs to rank Weblogs. Weblogs have their own special link structure which contains blog rolls, comments, and entry to entry links. We also employ Game Theoretic approach for Social Network Analysis. To do so, we formulate Nash Equilibrium for social network. |  |
|  |
Thesis: "Ontology Aligning Using Background Knowledge" |  |
Research Area: "News Minig" |  |
- Salman Mirghasemi, M.Sc. student
Thesis: "A semantic based framework for creation, maintenance and use of software knowledge model"
Despite of improvements in software engineering methods and tools, there is a gap between knowledge exists in mind of designers and developers and what is described by artifacts. Existence of a knowledge model which includes knowledge about artifacts, domain model and relations between them can be useful for better tool support and better software understanding.
Semantic Web is a new technology which creates a platform for knowledge representation and modeling. In our work we try to present a semantic based framework for creation and using a knowledge model in software development process.
|  |
Thesis: "Information Retrieval in Semantic Web Documents"
Semantic Web Documents (SWDs) are one of the information resources on the semantic web. They use semantic web languages and standards as a part of their contents. Based on distinctive characteristics of semantic web like providing semantic relations between SWDs and its formal infrastructure, improving and adapting traditional information retrieval techniques for SWDs is a challenging problem. We are interested in developing new models and adapting current techniques for gaining higher performance of information retrieval on the semantic web. |  |
Thesis: "A Framework for Brokering Semantic Web Services"
Currently, many web services are available on the web and their number is increasing exponentially. But in many cases these services can not fulfill users' needs. Service Composition means composing available web services for creating new value-added services. In this thesis, I plan to implement a service broker which receives users' requests and tries to match their requests with available web services. If no web service can fulfill the request it attemps to compose available services and create a new service which can satisfy user's request. In the composition process, the requester's context is also considered into account.
|  |
Thesis: "Ontology Aligning"
One of the core challenges for the semantic web is the aspect of centralization. Local structures can be modeled by ontologies. In order to support global communication and knowledge exchange, mechanisms have to be developed for the integrating of local systems. (Semautomatic alignment of ontologies is a core task to achieve interoperatability when two agents or services use different ontologies. In this thesis, I plan to implement a new framework and adapting existing approaches for gaining higher quality on ontology alignment. |  |
- Seyed H. HAERI (Hossein), B.Sc.
Research Area: Improving Ontology Alignment using Graph Theory
Given that Ontology Matching, and more generally Ontology Alignment, is in essence a Graph Homeomorphism problem, there is a vast and fertile ground for injecting Graph Theoretical techniques. I worked on the refinement techniques, and tried to harness the problem scale using two different schemas: consideration of the Concept Metric Spaces, and suiting the problem-specific intuitions for reducing the Typed Graph Homeomorphism problem.
I am currently a PhD student at the ULTRA Group, Heriot-Watt University, Edinburgh, UK.
|  |
|
|
|
|
|
|
|
|
People 