Hello, I'm using a dataset with unequally spaced time series and I'd want to know if there is in R some functions in order to calculate the autocorrelation function, because acf() in stats package cannot calculate it, because I have many missing data. And if so, is it necessary to organize dataset in ts structure or is it possible to consider it like a simple vector? Thanks in advance Fabrizio Consentino _______________________________________
Variogram with time series
4 messages · consentino@infinito.it, Brian Ripley
Hello, I'm using a dataset with unequally spaced time series and I'd want to know if there is in R some functions in order to calculate the autocorrelation function, because acf() in stats package cannot calculate it, because I have many missing data. And if so, is it necessary to organize dataset in ts structure or is it possible to consider it like a simple vector? Thanks in advance Fabrizio Consentino
15 days later
Hello, I'm using a dataset with unequally spaced time series and I'd want to know if there is in R some function in order to calculate the autocorrelation function, because acf() in stats package cannot calculate it, because I have many missing data, and data are not equally spaced. And if so, is it necessary to organize dataset in ts structure or is it possible to consider it like a simple vector? Thanks in advance Fabrizio Consentino
On Wed, 16 Mar 2005 consentino at infinito.it wrote:
I'm using a dataset with unequally spaced time series and I'd want to know if there is in R some function in order to calculate the autocorrelation function, because acf() in stats package cannot calculate it, because I have many missing data, and data are not equally spaced.
The autocorrelation function is a theoretical quantity: acf() estimates it, not calculates it. (Do read its help page carefully.) For `unequally spaced time series' you would need to define what you mean by the autocorrelation function, then find a credible estimator. (One way forward is to assume you have a sample of a continuous-time series, but only one.)
And if so, is it necessary to organize dataset in ts structure or is it possible to consider it like a simple vector?
None of the above! Depends what you want to do with it, but the structures offered by the packages tseries, zoo and its seem more suitable.
Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595