Re: AI-GEOSTATS: moving averages and trend

[José M. Blanco Moreno]

Dear Seba, As far as I know, the shape of the kernel is not that important. Any kernel will yield approximately the same results as long as the scaling parameters (bandwidth) are equivalent. There are some theoretical as well as practical reasons for choosing finite (compact) kernels - the practical reasons for big/huge problems become very important ones. José M. En/na seba ha escrit: Maybe I'm getting old...there was something wrong also in the past mail! Please consider this one. Sorry for the wrong mails. Dear list members I'm always fighting with the decomposition of trend and residuals or more generally I need to decompose my signal in high and low frequency variability. Without coming to wavelets techniques a simple way is to use moving window averages to obtain a smoothed signal and removing it from the original signal to obtain the residuals. In this way the size of the window let you choose at which level of detail to perform the analysis, i.e. smaller the window higher the frequency of the signal you want to study. But working with moving averages I realize (well, I know that this is not so a big new!) that performing moving window simply doing averages gives a smoothed signal that has some noise; differently if a use some kind of kernel function (also very simple such the one used by Grigov et al.  "geostatistical Mapping with continuos moving neighborhood", mathematical geology vol 36 no. 2, 2004 see page 273), things work really better (and using a kernel of that type is like calculating moving windows averages not on the original signal but iteratively on moving averages calculated in smaller windows....). Now my question is: which is the reason for choosing a specific shape of the kernel???? Sebastiano + + To post a message to the list, send it to ai-geost...@jrc.ec.europa.eu + To unsubscribe, send email to majordomo@ jrc.ec.europa.eu with no subject and "unsubscribe ai-geostats" in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/ -- --------------------------------------- José M. Blanco-Moreno Dept. de Biologia Vegetal (Botànica) Facultat de Biologia Universitat de Barcelona Av. Diagonal 645 08028 Barcelona SPAIN --------------------------------------- phone: (+34) 934 039 863 fax: (+34) 934 112 842 + + To post a message to the list, send it to ai-geost...@jrc.ec.europa.eu + To unsubscribe, send email to majordomo@ jrc.ec.europa.eu with no subject and "unsubscribe ai-geostats" in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list + As a general service to list users, please remember to post a summary of any useful responses to your questions. + Support to the forum can be found at http://www.ai-geostats.org/