Abstract The advent of big data characterized by 4V’s (i.e., volume, velocity, variety and veracity) imposes huge challenges to IT researchers and practitioners. In this paper, we review the basic concept of information granule (as studied in the field of granular computing), and discuss how this concept can be generalized to deal with new challenges imposed by big data. To accommodate 4 V’s, extensibility (the fundamental philosophy underlying extenics) is the desirable feature for information structure (represented by information granules), as elasticity the desirable feature for storage. We examine time-varying information granules, and the newly proposed concept of dynamic granule, to explain the need for considering soft information granules. We discuss how methodology offered by extenics can define soft boundaries for soft information granules, so that extensibility can play an effective role in dealing with unconventional features of big data. Examples are provided to illustrate our idea.
[1]
Yiyu Yao,et al.
Interpreting Concept Learning in Cognitive Informatics and Granular Computing
,
2009,
IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).
[2]
Curtis E. Dyreson,et al.
Efficiently Supported Temporal Granularities
,
2000,
IEEE Trans. Knowl. Data Eng..
[3]
Daniel F. Leite,et al.
Evolving fuzzy granular modeling from nonstationary fuzzy data streams
,
2012,
Evol. Syst..
[4]
Lotfi A. Zadeh,et al.
Fuzzy sets and information granularity
,
1996
.
[5]
Zhengxin Chen,et al.
Toward Extenics-Based Innovation Model on Intelligent Knowledge Management
,
2014
.
[6]
Fernando Gomide,et al.
Evolving granular analytics for interval time series forecasting
,
2016,
Granular Computing.