Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1149526 | Journal of Statistical Planning and Inference | 2010 | 12 Pages |
Abstract
In modern statistical practice, it is increasingly common to observe a set of curves or images, often measured with noise, and to use these as the basis of analysis (functional data analysis). We consider a functional data model consisting of measurement error and functional random effects motivated by data from a study of human vision. By transforming the data into the wavelet domain we are able to exploit the expected sparse representation of the underlying function and the mechanism generating the random effects. We propose simple fitting procedures and illustrate the methods on the vision data.
Related Topics
Physical Sciences and Engineering
Mathematics
Applied Mathematics
Authors
R. Todd Ogden, Ernest Greene,