Synopses for massive data: samples, histograms, wavelets, sketches (Q2903466)

In the last decades the amount of data stored in databases has grown very fast. A massive data set is a collection of complex and very large data, having multiple dimensions. Processing this data is a challenge for many researchers. Challenges include collecting, storing, querying, sharing, analyzing and visualizing the data.NEWLINENEWLINESuch massive data sets are stored on disks and querying is slow because the access to disks is slower than accessing direct memories. In order to reduce query execution time, the synopses data can be used.NEWLINENEWLINENEWLINE This paper provides the principles and methods to build approximate massive data synopses. It is structured into 6 chapters and a bibliography. The first chapter reveals the theoretical concepts presented in the paper. The following four chapters present the main types of synopses: samples, histograms, wavelets and sketches. Each of these chapters treats these notions gradually starting from their theoretical definition and subsequently presenting methods on how to obtain synopses and to perform the approximate query processing on these data.NEWLINENEWLINENEWLINE Finally, in the last chapter the authors present their conclusions and future research directions. The paper is clearly written and structured. The readers can achieve a good understanding of the theoretical concepts, which are very well exemplified.NEWLINENEWLINENEWLINE The paper is a valuable reference for the researchers interested in massive data, and is thus useful both for teaching and research.