decevf                package:pastecs                R Documentation

_T_i_m_e _s_e_r_i_e_s _d_e_c_o_m_p_o_s_i_t_i_o_n _u_s_i_n_g _e_i_g_e_n_v_e_c_t_o_r _f_i_l_t_e_r_i_n_g (_E_V_F)

_D_e_s_c_r_i_p_t_i_o_n:

     The eigenvector filtering decomposes the signal by applying a
     principal component analysis (PCA) on the original signal and a
     certain number of copies of it incrementally lagged, collected in
     a multivariate matrix. Reconstructing the signal using only the
     most representative eigenvectors allows filtering it.

_U_s_a_g_e:

     decevf(x, type="additive", lag=5, axes=1:2)

_A_r_g_u_m_e_n_t_s:

       x: a regular time series ('rts' under S+ and 'ts' under R) 

    type: the type of model, either `type="additive"' (by default), or
          `type="multiplicative"' 

     lag: The maximum lag used. A PCA is run on the matrix constituted
          by vectors lagged from 0 to `lag'. The defaulf value is 5,
          but a value corresponding to no significant autocorrelation,
          on basis of examination of the autocorrelation plot obtained
          by `acf' in the library 'ts' should be used (Lag at first
          time the autocorrelation curve crosses significance lines
          multiplied by the frequency of the series). 

    axes: The principal axes to use to reconstruct the filtered signal.
          For instance, to use axes 2 and 4, use `axes=c(2,4)'. By
          default, the first two axes are considered (`axes=1:2') 

_V_a_l_u_e:

     a 'tsd' object

_A_u_t_h_o_r(_s):

     Frdric Ibanez (ibanez@obs-vlfr.fr), Philippe Grosjean
     (phgrosjean@sciviews.org)

_R_e_f_e_r_e_n_c_e_s:

     Colebrook, J.M., 1978. Continuous plankton records: zooplankton
     and environment, North-East Atlantic and North Sea 1948-1975.
     Oceanologica Acta, 1:9-23.

     Ibanez, F. & J.C. Dauvin, 1988. Long-term changes (1977-1987) on a
     muddy fine sand Abra alba - Melinna palmate population community
     from the Western English Channel. J. Mar. Prog. Ser., 49:65-81.

     Ibanez, F., 1991. Treatment of data deriving from the COST 647
     project on coastal benthic ecology: The within-site analysis. In:
     B. Keegan (ed.) Space and time series data analysis in coastal
     benthic ecology. Pp. 5-43.

     Ibanez, F. & M. Etienne, 1992. Le filtrage des sries
     chronologiques par l'analyse en composantes principales de
     processus (ACPP). J. Rech. Ocanogr., 16:27-33.

     Ibanez, F., J.C. Dauvin & M. Etienne, 1993. Comparaison des
     volutions  long-terme (1977-1990) de deux peuplements
     macrobenthiques de la Baie de Morlaix (Manche Occidentale):
     relations avec les facteurs hydroclimatiques. J. Exp. Mar. Biol.
     Ecol., 169:181-214.

_S_e_e _A_l_s_o:

     `tsd', `tseries', `decaverage', `deccensus', `decmedian', 
     `decdiff', `decreg', `decloess'

_E_x_a_m_p_l_e_s:

     data(releve)
     melo.regy <- regul(releve$Day, releve$Melosul, xmin=9, n=87,
             units="daystoyears", frequency=24, tol=2.2, methods="linear",
             datemin="21/03/1989", dateformat="d/m/Y")
     melo.ts <- tseries(melo.regy)
     library(ts)
     acf(melo.ts)
     # Autocorrelation is not significant after 0.16
     # That corresponds to a lag of 0.16*24=4 (frequency=24)
     melo.evf <- decevf(melo.ts, lag=4, axes=1)
     plot(melo.evf, col=c(1, 4, 2))
     # A superposed graph is better in the present case
     plot(melo.evf, col=c(1, 4), xlab="stations", stack=FALSE, resid=FALSE,
             lpos=c(0, 60000))

