Re: Random Coeffecients Model
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Random Coeffecients Model




Dear Terry,

I havn't checked the model in detail but there is a problem with
the specification of the G structure at the bottom.

 pot 2   # define 2 variance structures for Pots
 48 0 I  # independence between Pots
 2 0 CORR .329 82.8914 15.8124 # but corr{intercept.slope} within Pots

SHOULD BE

pot 2   # define 2 variance structures for Pots
2 0 CORR .329 82.8914 15.8124 # but corr{intercept.slope} within Pots
48 0 I  # independence between Pots

because the 96 effects to which this structure
is applied (pot and pot.vtime)

are arranged as 2 sets of 48, not 48 sets of 2.


Two minor comments:
1.  you could have used  dev(vtime)  in place of  ftime

2.  you are right that the multivariate input format
used in the random coefficients example on page 143/144
is not suitable for unequally spaced times (the more usual situation).
Note however that the G structure definition there is consistent
with what I have specified above.

Best wishes for Christmas and the new year.

Arthur

> 
> Dear ASREMLers,
> 
> Would someone please care to run their eye(s) over my code below ----- I am
> fitting a full model of terms, hoping to be correctly including a random
> slopes
> and intercepts model, correlated for each unit.
> 
> At 18 (un-equally spaced) times, records were taken of the water usage of
> plants in 48 pots (made up of 3 reps of 4 species by 2 water regimes by 2
> fertilizer rates).  Since my time intervals are not consistent, I cannot see
> how to use the example on pages 143/144  of the Oct 1998 manual.
> 
> Thanks for your time.
> Cheers,  Terry 
> ========================================================
> Bill Johnston's water use data: 4 sp X 2 Water X 2 Fert
> # 18 times * 48 pots = 864 units; time 1 data (26 Sept) dropped
> #
>    species    4 !A   # Consol, 4650, Lucerne, Phalaris
>    water      2 !A   # Low, High
>    fertil     2 !A   # Low, High
>    treat     16      # Factor coding for the above 3 treatments
>    rep        3      # Factor for reps
>    pot       48      # Factor for 3rep * 4 sp * 2 water * 2 fertilizer
>    vtime        !-340.778 !/25.879   # Variate days {291...379}
>    fdate     18 !A   # Factor of 18 dates: 18-Oct, 30-Oct, .......
>    ftime     18      # Factor of 18 dates: 1,2,3,...,18
>    usage             # cumulative water usage data
> #
> d:\1_data\johnston\wateruse\BJ_treat\wuetrial.dat !skip 1 !maxit 30
> #
> # Fit a random coefficients model
> usage  ~ mu species*water*fertil ,
>    vtime vtime.spe vtime.wat vtime.fer,
>    vtime.spe.wat vtime.spe.fer vtime.wat.fer,
>    vtime.spe.wat.fer,
>    !r pot                  82.8914, # random intercepts for pots
>    pot.vtime               15.8124, # random slopes for pots
>    spl(vtime)               0.00001,
>    spe.spl(vtime)           1.5492,
>    wat.spl(vtime)          36.0471,
>    fer.spl(vtime)           0.01632,
>    spe.wat.spl(vtime)       0.07648,
>    spe.fer.spl(vtime)       0.00001,
>    wat.fer.spl(vtime)       0.009266,
>    spe.wat.fer.spl(vtime)   0.00001,
>    ftime                   17.0388,  # random lack of fit for time
>    ftime.treat              0.062343,# departures from smoothness for treats
>    pot.spl(vtime)           1.14104  # pick up curvature for each pot
>         
> 0 0 1   # there is 1 G header line
> pot 2   # define 2 variance structures for Pots
> 48 0 I  # independence between Pots
> 2 0 CORR .329 82.8914 15.8124 # but corr{intercept.slope} within Pots
> ====================================================================== 
> *******************************************************************
> Terry Koen							EMail: 	
koentb@ozemail.com.au
> Research Centre						Phone: 	(02) 
6342 1811 (w)
> Dept Land and Water Conservation				(02) 6341 1368 
(h)
> PO Box 445, Cowra, NSW 2794,		Fax:	(02) 6342 4551
> Australia
> *******************************************************************
> --
> Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Arthur Gilmour PhD                    email: Arthur.Gilmour@agric.nsw.gov.au
Senior Research Scientist (Biometrics)                 fax: <61> 2 6391 3899
NSW Agriculture                                             <61> 2 6391 3922
Orange Agricultural Institute               telephone work: <61> 2 6391 3815
Forest Rd, ORANGE, 2800, AUSTRALIA                    home: <61> 2 6362 0046

ASREML is still free by anonymous ftp from pub/aar on ftp.res.bbsrc.ac.uk
    or point your web browser at ftp://ftp.res.bbsrc.ac.uk/pub/aar/ 

To join the asreml discussion list, send the message  
     subscribe
to  asreml-request@chiswick.anprod.CSIRO.au

The address for messages to the list is asreml@chiswick.anprod.CSIRO.au

Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml

                        <> <> <> <> <> <> <>
"For the wages of sin is death,
  but the gift of God is eternal life in Christ Jesus our Lord." Romans 6:23
<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>


--
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml