Barnard, EVan der Walt, Christiaan MDavel, MVan Heerden, CSenekal, FPNaidoo, T2012-02-152012-02-152009-11Barnard, E, Van der Walt, C, Davel, M et al. Learning structured representations of data. 20th Annual Symposium of the Pattern Recognition Association of South Africa (PRASA), Stellenbosch, South Africa, 30 November-01 December 2009, pp 1-6978-0-7992-2356-9http://www.prasa.org/proceedings/2009/prasa09-01.pdfhttp://hdl.handle.net/10204/557020th Annual Symposium of the Pattern Recognition Association of South Africa (PRASA), Stellenbosch, South Africa, 30 November-01 December 2009Bayesian networks have shown themselves to be useful tools for the analysis and modelling of large data sets. However, their complete generality leads to computational and modelling complexities that have limited their applicability. We propose an approach to simplify and constrain Bayesian networks that strikes a more useful compromise between generality and tractability. These constrained graphical will allow us to build computationally tractable models for large high-dimensional data sets. We also describe examples of data sets drawn from image and speech processing on which can (1) further explore this constrained set of graphical models, and (2) analyse their performance as a general-purpose statistical data analysis tool.enData setsData analysisGeneralityTractabilityBayesian networksConference PresentationBarnard, E., Van der Walt, C. M., Davel, M., Van Heerden, C., Senekal, F., & Naidoo, T. (2009). Learning structured representations of data. PRASA. http://hdl.handle.net/10204/5570Barnard, E, Christiaan M Van der Walt, M Davel, C Van Heerden, FP Senekal, and T Naidoo. "Learning structured representations of data." (2009): http://hdl.handle.net/10204/5570Barnard E, Van der Walt CM, Davel M, Van Heerden C, Senekal F, Naidoo T, Learning structured representations of data; PRASA; 2009. http://hdl.handle.net/10204/5570 .TY - Conference Presentation AU - Barnard, E AU - Van der Walt, Christiaan M AU - Davel, M AU - Van Heerden, C AU - Senekal, FP AU - Naidoo, T AB - Bayesian networks have shown themselves to be useful tools for the analysis and modelling of large data sets. However, their complete generality leads to computational and modelling complexities that have limited their applicability. We propose an approach to simplify and constrain Bayesian networks that strikes a more useful compromise between generality and tractability. These constrained graphical will allow us to build computationally tractable models for large high-dimensional data sets. We also describe examples of data sets drawn from image and speech processing on which can (1) further explore this constrained set of graphical models, and (2) analyse their performance as a general-purpose statistical data analysis tool. DA - 2009-11 DB - ResearchSpace DP - CSIR KW - Data sets KW - Data analysis KW - Generality KW - Tractability KW - Bayesian networks LK - https://researchspace.csir.co.za PY - 2009 SM - 978-0-7992-2356-9 T1 - Learning structured representations of data TI - Learning structured representations of data UR - http://hdl.handle.net/10204/5570 ER -