 |
 |
The following course will be of interest to anyone interested in "Wavelets". Please register early to guarantee a place. Course: Wavelets in Statistics Wednesday 26th April 2006 A One-Day Workshop at the University of Bristol by Piotr Fryzlewicz Cost: £55 includes tea, coffee, lunch and course materials (discounts available for participants from same university/college/workplace - £100 for two, £140 for three) Registration: Contact Brad Payne on 0115 8488410 or email [log in to unmask] <mailto:[log in to unmask]> (Please state if you have any special dietary requirements) Synopsis Although wavelets were "discovered" by statisticians relatively recently (in the early nineties), they are now an established tool for tasks such as data smoothing, density estimation, or image compression. Wavelets can be informally described as localised, oscillatory functions, designed to have several attractive properties not enjoyed by "big waves" - sines and cosines. They are usually arranged in multiscale bases and are thus capable of extracting local information from the data at a range of scales. In this course, I will first motivate wavelets by showing how they can be used for data smoothing in the "function + iid Gaussian noise" setup. I will then outline the mathematical construction of wavelets, mainly using the example of Haar wavelets, which form the simplest wavelet family. I will then describe the Discrete Wavelet Transform, and finally discuss in more detail the application of wavelets to data smoothing. The practical part of the course will be a hands-on tutorial on "WaveThresh", a free software package for wavelet analyses for use in the freeware R environment. As an illustration, we will look at the wavelet smoothing of the famous "Wolf's sunspot numbers" data set. Programme (provisional) 09.45 - 10.15 Registration and Welcome 10.15 - 11.15 Session 1: Wavelets - motivation and mathematical construction 11.15 - 11.30 Morning Coffee 11.30 - 12.30 Session 2: Discrete Wavelet Transform 12.30 - 13.30 Lunch 13.30 - 14.30 Session 3: Data smoothing via wavelets 14.30 - 14.45 Afternoon Tea 14.45 - 15.45 Session 4: WaveThresh - hands-on tutorial and real data example 15.45 - 16.15 Discussion and Close Biopic Piotr Fryzlewicz is a lecturer in statistics in the Department of Mathematics at the University of Bristol. His research interests are in multiscale methods in statistics (including the wavelet-Fisz methodology), nonparametric regression and time series. One of his ongoing projects is on "locally stationary financial time series models" and is funded by the Nuffield Foundation. This email is intended solely for the addressee. It may contain private and confidential information. If you are not the intended addressee, please take no action based on it nor show a copy to anyone. In this case, please reply to this email to highlight the error. Opinions and information in this email that do not relate to the official business of Nottingham Trent University shall be understood as neither given nor endorsed by the University. Nottingham Trent University has taken steps to ensure that this email and any attachments are virus-free, but we do advise that the recipient should check that the email and its attachments are actually virus free. This is in keeping with good computing practice.