Abstract:In this study, we are concerned with a two-objective development of information granules completed on a basis of numeric data. The first goal of this design concerns revealing and representing a structure in a data set. As such it is very much oriented towards coping with the underlying relational aspects of the experimental data. The second goal deals with a formation of a mapping between information granules constructed in two spaces (thus it concentrates on the directional aspect of information granulation). The quality of the mapping is directly affected by the information granules over which it operates, so in essence we are interested in the granules that not only reflect the data but also contribute to the performance of such a mapping. The optimization of information granules is realized through a collaboration occurring at the level of the data and the mapping between the data sets. The operational facet of the problem is cast in the realm of fuzzy clustering. As the standard techniques of fuzzy clustering (including a well-known approach of FCM) are aimed exclusively at the first objective identified above, we augment them in order to accomplish sound mapping properties between the granules. This leads to a generalized version of the FCM (and any other clustering technique for this matter). We propose a generalized version of the objective function that includes an additional collaboration component to make the formed information granules in rapport with the mapping requirements (that comes with a directional component captured by the information granules). The additive form of the objective function with a modifiable component of collaborative activities makes it possible to express a suitable level of collaboration and to avoid a phenomenon of potential competition in the case of incompatible structures and the associated mapping. The logic-based type of the mapping (that invokes the use of fuzzy relational equations) comes as a consequence of the logic framework of information granules. A complete optimization method is provided and illustrated with several numeral studies.

Granular Worlds: Representation and Communication Problems

Communications Between Granular Worlds: Fundamentals

GRANULAR COMPUTING: CONCEPTS AND ALGORITHMIC DEVELOPMENTS

Granular Computing : an Introduction

Information Granules and Granular Models: Selected Design Investigations

Granular Computing: Fundamentals and System Modeling

From Numeric Models to Granular System Modeling

Granular Computing for System Modeling

Managing Granular Information in the Development of Human-Centric Systems

Human Centricity and Perception-Based Perspective and Their Centrality to the Agenda of Granular Computing

Concepts and Design of Granular Models: Emerging Constructs of Computational Intelligence

Granular Computing as a Framework of System Modeling

Information Granules and Their Use in Schemes of Knowledge Management

Granular Computing in Fuzzy Modeling and Data Mining

Relations of Granular Worlds

Generalizations of Information Granules

Granular Computing as an Emerging Paradigm of Information Processing

Granular Computing: Perspectives and Challenges

Regranulation: A granular algorithm enabling communication between granular worlds

Directional Models of Granular Communication

The Principle Of Justifiable Granularity And An Optimization Of Information Granularity Allocation As Fundamentals Of Granular Computing