Analysis of spatiotemporal data relationship using information granules
- PDF / 1,239,582 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 40 Downloads / 183 Views
ORIGINAL ARTICLE
Analysis of spatiotemporal data relationship using information granules Mingli Song1 • Wenqian Shang1 • Lidong Wang2 • Witold Pedrycz3
Received: 30 November 2014 / Accepted: 4 June 2015 Springer-Verlag Berlin Heidelberg 2015
Abstract Data analysis especially data with space and time feature in a human-centric way requires interpretable representation of data. With this motivation, we present a granular way of data analysis in which the data and the relationships therein are described through a collection of sets or fuzzy sets (information granules). In this paper, data are described by semantically meaningful descriptors-information granules over the space and time domain. The design process is guided by information granulation and degranulation. Thus a performance index used to obtain the best combination of information granules becomes a crucial issue. The effectiveness of the algorithm is demonstrated by experiments on two kinds of synthetic data and data from Alberta agriculture website. Keywords Spatiotemporal data Information granules FCM Granular descriptor
& Mingli Song [email protected]; [email protected] Wenqian Shang [email protected] Lidong Wang [email protected] Witold Pedrycz [email protected] 1
School of Computer Science, Communication University of China, Beijing 100024, China
2
Department of Mathematics, Dalian Maritime University, Dalian 116026, China
3
Department of Electrical and Computer Engineering, University of Alberta, Edmonton T6R 2G7, Canada
1 Introduction Spatiotemporal data become visible in various application fields such as marketing, land use, economics, engineering, industry, multimedia and so on. We have been witnessing a slew of approaches originating from studies developed within the setting of analysis and processing of spatiotemporal data. Some studies focus on mining knowledge from the data [1, 2], some try to build a model using spatiotemporal data [3], and some others determine the relationship of spatiotemporal data [4]. The need for a concise, highly interpretable, and accurate descriptors of data is highly visible so that such descriptions reveal and describe an essence of the main relationships and associations among variables of the systems. Information granules, as the name itself stipulates, are collections of entities, usually originating at the numeric level, that are arranged together due to their similarity, functional adjacency, and coherency or alike. Information granules are seen everywhere: spatial granulation like image processing and geographic information system (GIS), temporal granulation like cultural, legal, business orientation of the designer and so on. They are central to processes of abstraction which guide our intellectual pursuits. In this paper, we concentrate on the study of representation of spatiotemporal data with information granules which will help the subsequent analysis of spatiotemporal data. The spatial and temporal data can be described at various levels of abstraction. In fact, this lev
Data Loading...
