MANAGEMENT OF BIG DATA IN SCIENTIFIC COMPUTING

Abstract

The recent developments in computer processing has led to the generation of significant amount of data. However, the post-processing of this large amount of data poses significant challenges for the scientific community. A particular issues is the data interpretation, particularly regarding the segregation of valuable data from arbitrary data. Moreover, data classification in subgroups poses further challenges. It has been largely accepted that the analysis of these data sets may be cumbersome. Therefore, efficient and accurate data mining models would enable the analysis if large data-sets. In this research we propose a computational model based on the k-means algorithm, identified as fuzzy clustering algorithm. The use of the fuzzy clustering algorithm reduces the computational time by 72%, when compared with regular computational approaches.

Keywords

• data management
• numerical modeling
• adaptive mesh refinement
• k-means algorithm