Detail Longitudinal Statistical Modelling on The Grid   Oleh:  Crouchley, Rob ; Allan, Rob Jenis: Article from Books - Reference Dalam koleksi: The SAGE Handbook of Online Research Methods, page 471-490. Topik: Longitudinal Statistical Modelling; Computational Power; e-Science Technology and The Grid; Collaboration in Virtual Communities Ketersediaan Perpustakaan PKPM Nomor Panggil: 301.150.1 SAG 7 Non-tandon: tidak ada Tandon: 1  Lihat Detail Induk Isi artikel We believe that the social science will undertake a gradual shift to research practices that involve transparent access to large-scale information , data and software and will use high performance computers and collaboration resources to tackle more complex and challenging problems than at present. In this chapter we illustrate this by describing the motivation and some solutions being adopted by innovators involved in quantitative social science studies of individual behaviour. The rate of social science engagement with the Grid paradigm will vary, however, depending on the specific research challenges/interests and technical ability of the various groups involved. Some will naturally wait while the large-scale software houses like Microsoft [1; references in brackets are to the URL link list, end of chapter], IBM [2] and SAS [3] evolve the tools and technologies they use. Others, who have already reached the limit of what they can do with the current technology, are pressing for innovation, and some are already doing it for themselves. Whilst the number of social scientists in this second group is currently a very small fraction of the social science research community, we estimate that there are many more researchers who could use some of this technology in the medium to longer term. This is dependent, however, on the pathfinders both showing the way and illustrating the potential benefits from grid computing. We are aware of several instances in which the substantive demands of our research require the use of datasets that are too large to be managed or manipulated, and/or the complexity of the statistical models we need to estimate is too demanding, for the current range of PCs. We will show what has been achieved so far to tackle this using our own Open Source software, Sabre and GROWL. This provides access, from a familiar environmentsuch as the R statistical language, to estimate statistical models of complex social processes on Grid resources much more quickly than by using conventional applications. Opini Anda Klik untuk menuliskan opini Anda tentang koleksi ini! Kembali

Copyright © 2006, 2007 Unika Atma Jaya, all rights reserved