• home
• current students
• graduate
• events

Presenter: Dr. Witold Pedrycz - Dept. of Electrical & Computer Engineering, University of Alberta
Supervisor:

Date: Fri, March 18, 2005
Time: 14:30:00 - 15:30:00
Place: COR A125

ABSTRACT

Abstract

Human-centric systems and human-centric computing (HC 2) form a new and already visible trend in modern information technology. The primary objective of the HC 2 is to make computers adjust to people by being more natural and intuitive to use or seamlessly integrated with the existing environment. Various pursuits along the line of e-society including intelligent housing, semantic web, e-health, e-commerce, intelligent data analysis, and wearable hardware are examples of the tendency existing in the development of HC 2 systems. In this presentation, we will concentrate on the underlying concepts, role and selected algorithms of knowledge-based clustering being regarded as an essential technology for human-centric systems. Starting with a cornerstone notion of information granules (whose construction is inherently supported by various mechanisms of clustering), we discuss a role they play in the realization of the HC 2 and present formal frameworks they usually entail. We elaborate on various ways of incorporating knowledge-based guidance into the predominantly data-driven clustering mechanisms when realized by means of fuzzy objective function-based clustering. In particular, we present mechanisms of partial supervision, context-based clustering, and proximity-oriented clustering. Collaborative clustering is discussed as an interesting vehicle of forming consensus between structures and relationships existing in various sources of data. Shown are also some pertinent applications of the environment of knowledge-based clustering to Web-based data and organization of digital images.