2002 - Present

WIDIT Research

KoDDAS Project: 2019-present, Co-PI

Korean Diaspora Digital Archive System (KoDDAS) is a 6-year National Research Foundation project to develop an integrated database system that utilizes augumented-intelligence and faceted classification. The purpose of KoDDAS is to provide a digital data management infrastructure for aggregating, organizing, and disseminating Korean Diaspora data for research communities as well as the general public. In that regard, KoDDAS extends the conventional archival functions of collection and preservation to focus on information dissemination and data reuse. To promote data reuse, KoDDAS will implement an integrated online database system that contains not only the raw data, but also the contextual and structural metadata that can facilitate the understanding and analysis of Korean Diaspora data.

TREC Project: 2002-2009, Project Director

TREC project investigates cutting-edge information retrieval strategies that accommodate a variety of data environments and search tasks. TREC project is responsible for developing core components of the WIDIT system architecture, which are utilized in other WIDIT projects, as well as TREC experiments.

CiteSearch Project: 2005-present, Project Director

CiteSearch project, a component of VCoB, develops a multi-faceted fusion approach to information quality assessment that employs a range of citation-based methods to analyze data from multiple sources. The CiteSearch system currently consists of the sf-vcob component, which harvests and analyzes citation data of faculty publications, and the mini-vcob component, which collects and organizes scholarly publications and associated citation data on a given topic.

TREC Project: 2002-2009, Project Director

TREC project investigates cutting-edge information retrieval strategies that accommodate a variety of data environments and search tasks. TREC project is responsible for developing core components of the WIDIT system architecture, which are utilized in other WIDIT projects, as well as TREC experiments.

VCoB Project: 2003-2009, Project Director

Virtual Collection Builder (VCoB) project aims to develop an adaptive and interactive agent for building and maintaining a virtual collection of Web documents. VCoB will employ a wide array of methods from content analysis, citation analysis, machine learning, and information retrieval to collect, organize, and maintain a searchable and browsable collection of documents that are custom-taylored to the needs and preferences of individual users.

CSKD Project: 2003 - 2009, Project Co-Director with E.K. Jacob

The Classification-based Search and Knowledge Discovery (CSKD) research explores methods of leveraging both the ontological and link-structural knowledge embedded in classified corpora of Web documents for searching and organizing the Web. CSKD is a multi-dimensional project that entails investigations in such area as machine learning, classification, clustering, link analysis, and fusion.

JiTTDL Project: 2004 - 2007, Project Co-Director with E.K. Jacob

Just-in-Time Teaching Digital Library (JiTTDL) is an NSF-funded project at US Air Force Academy to develop the digital library for Justin-in-Time pedagogical resources, where CSKD-based approaches were employed to construct a prototype digital library system.

DGov Project: 2004 - 2006, Project Co-Director with L. Meho

Digital Government (DGov) project investigates an efficient and effective approach to organizing the U.S. Government websites that can facilitate the access and enhance the retrieval of government information.i The DGoV project became inactive when a major government initiative to restructure FirstGov.gov to USA.gov began in 2007.

1999 - 2002

Sitemap Tree Project

• A dynamic tree manipulation application allowing creation and maintenance of a hierarchical tree structure, which can represent various entities from a file system (e.g. Windows Explorer) to a hyperlink-based map of a given Web site.
• http://ils.unc.edu/yangk/sitemaptree.htm

1998

IRISWeb project under Chancellor’s Instructional Technology Grant

• IRISWeb is a work-in-progress prototype of the next-generation web search engine that will incorporate the features of dynamic indexing, meta-search, relevance-feedback, and dynamic collection development.
• Currently implemented in Java, C++, C, and Perl.
• http://ils.unc.edu/iris/

1997

IRISWeb Indexing Project

• IRIS Web Indexing module is comprised of two main components: the Web Crawler/Indexer application which indexes a target URL and its embedded links, and Sitemap Display applet, which displays the indexed pages in a directory tree structure.
• Implemented in Java.
• http://ils.unc.edu/iris/sitemap/mksmap.htm

1996

Nice Stemmer Project

• Nice Stemmer implements 4 stemmer algorithms, any of which can be executed by the user according to his/her need.
• Implemented in C++.
• http://ils.unc.edu/yangk/nstem.htm

IRISRuby Project

• IRISRuby is the first incarnation of IRIS. It can index and search documents stored on an intranet server.
• Implemented in Perl.
• http://ils.unc.edu/inls300/iris.html