Past and Current Work

 

Summary of current work:

Work during my PhD primarily focused on the application of wavelets to time series and spatial data. Motivated by the work of Nason, von Sachs ∓ Kroisandt I extended the LSW modelling approach to two-dimensions. Such methodolgy naturally has various texture analysis applications, details of which can be found in the forthcoming publication with G.P. Nason (see below). The estimation schemes assoicated with this new modelling approach have been implemented in S-Plus. For further details, please visit the LS2W-toolbox web-site.

Some of the projects I am currently working on include:

  • Simulation of an textured images using this 2-D locally stationary model.

  • Estimation of the local autocovariance of locally stationary wavelet processes.

  • Applications of wavelet methods to change-point problems.

  • Development of multiple classifier systems.

    • Publications

    Eckley, I. A. (2001) Wavelet methods for time series and spatial data , PhD Thesis, University of Bristol. [Abstract], [PDF]

    Eckley, I. A. and Nason, G. P. (2001) The inner product matrix of discrete autocorrelation wavelets: efficient computation and application, (In submission) [Abstract], [PDF-330k], [PostScript-784k]

    McDonald, R. A., Eckley, I. A. and Hand, D. J. (2002) A multiclass extension to Brownboost (To appear in International Journal of Pattern Recognition and Artificial Intelligence) [Abstract], [PDF], [PostScript]

    McDonald, R. A., Eckley, I. A., Jonathan, P. and Hand, D. J. (2002) Boosted aggregate models in commerce and industry: an introduction and case study (In submission) [Abstract], [Gzipped PDF], [Gzipped PostScript]

    McDonald, R. A., Hand, D. J. and Eckley, I. A. (2003) An empirical comparison of three boosting algorithms on real data sets with artificial class noise 35-44 Proceedings of Multiple Classifier Systems 2003, LNCIS Volume 2709, Springer-Verlag (Heidelberg) [Abstract]

    Eckley, I.A., Nason, G. P. and Treloar, R. L. (2003) Locally stationary wavelet fields with application to the modelling and analysis of image texture (In submission) [Abstract], [PDF]

    Recent Talks and Presentations


    1. Looking at pictures from a different perspective. Research Students Conference, Newcastle, March 26, 2001.

    2. Wavelets and texture analysis. Young Statisticians Meeting, Liverpool, April 9-10, 2001.

    3. Locally stationary wavelet processes: a review of recent work (joint with G. P. Nason). NDTSA, Luminy, April 26 2001.

    4. Locally stationary processes and their application to texture analysis (joint with G. P. Nason). Joint Statistical Meetings, Atlanta, August 6, 2001.

    5. Local autocovariance estimation. International conference of the Royal Statistical Society, Plymouth, August 6, 2002.

    6. Boosting classification performance. Young Statisticians Meeting, Cambridge, April 10, 2003.
     

     

    Collaborators:


    David Hand, Philip Jonathan, Ross McDonald, Guy Nason, Rob Treloar

    HomeLast modified: Thu Mar 29 17:03:24 BST