BROWSE
Detail
Conference Paper
Hybrid Clustering Framework for Multi-dimensional Array Data
Cited - time in 
Share 
Share
- Authors
- Hyeon Park, Dae-Heon Park, Eun-Ju Lee, Se-Han Kim
- Issue Date
- 2015-10
- Citation
- International Conference on Big Data Applications and Services (BigDAS) 2015, pp.1-4
- Language
- English
- Type
- Conference Paper
- Abstract
- As the satellite imagery containing multi-dimensional array data is currently used for analysis of various applications, the frameworks to analyze that sort of scientific data have been introduced. To process the scientific data like the satellite imagery there are some restrictions: for the analysis of large-scale data the aggregated data would be stored in specified data formats, for the time-series analysis of the huge size the specified file system would be needed as the data is rapidly increased, and so on. Although Hadoop framework which is big data computing platform is popular to process the big data it is not feasible to handle the scientific data. It does not support to process the data in different scientific formats. On the other hand, though SciDB is the data management system to mainly process large-scale array data, it is not appropriate to analyze the scalable data of the time series. In this paper, we propose hybrid clustering framework, which is to process the scientific data composed of the multidimensional arrays with time series. The proposed framework would address the issues to provide the framework both processing array-based scientific data and handling ever-increasing data at the same time.
- KSP Keywords
- Aggregated data, Array data, Computing platform, Data Management System, Hadoop Framework, Large-scale Data, Large-scale array, Multi-dimensional array, Satellite imagery, Scientific data, Time Series Analysis(TSA)