Greetings R Help Group, How does one effect a multiphasic logistic growth model with 4 phases (e.g. Koops 1986; Weigel, Craig, Bidwell and Bates 1992; Grossman and Koops 2003) with R. Before writing to the group, the R help archives were searched, the web was searched with Google, Venables and Ripley 2002 was consulted, Pinheiro and Bates 2000 was consulted, Bates and Watts 2007 was bought and consulted, ETC. but to no avail. I have not written to any other group with respect to this problem. The following data are offered as an example of the type of problem I am dealing with and are average daily goat milk production in ml. for 14 weeks. NoC is the goat number. PL1 is the production during the first week; PL2 is the production during the second week, etc. NoC PL1 PL2 PL3 PL4 PL5 PL6 PL7 PL8 PL9 PL10 PL11 PL12 PL13 PL14 52 490 950 800 900 850 850 750 610 640 900 980 890 890 910 48 970 1020 980 740 1050 970 850 790 900 920 1120 1120 1030 1300 54 1450 1420 1430 1120 1330 1230 1030 1170 1350 1530 1490 1500 1310 910 60 620 1250 1100 1150 780 930 990 940 760 730 790 1050 840 850 58 1370 1350 1200 1300 1350 1310 1070 910 1010 1300 1110 1070 990 660 42 1200 920 1150 850 720 630 630 710 850 810 930 980 820 1570 44 1200 1560 1650 1600 1450 1600 1160 1010 1440 1450 1530 1500 1550 850 47 1000 1100 1000 870 760 900 820 865 910 820 930 900 1130 1070 50 1300 1150 1200 1030 1070 970 860 900 950 1190 1250 1130 800 1400 56 950 1250 1200 1280 1220 1155 840 1016 1370 1220 1570 1520 1500 1150 1 870 1250 1160 1270 1200 1410 1110 1008 970 1130 1490 1330 1320 820 3 1000 1100 1200 1120 1250 980 750 890 1050 1160 1340 1210 1150 760 4 551 760 550 580 540 620 550 520 470 720 680 790 750 1230 5 810 1100 820 950 930 830 850 650 810 1070 1120 1300 1040 1320 6 800 1000 620 850 750 670 660 620 600 610 760 900 758 1070 7 720 830 1120 1050 820 820 850 810 800 750 780 940 1050 1310 8 950 1550 1560 1500 1230 1330 1150 1005 1020 1200 1440 1400 1290 1080 10 660 850 1100 980 1070 1100 870 790 880 950 1000 1210 1050 1220 It seems to me that it should be possible to effect the modeling process with nlme. Any suggestions and or recommendations would be greatly appreciated. Peter B. Douglas M. Bates and Donald G. Watts. 2007. Nonlinear Regression Analysis and Its Applications. Wiley Series in Probability and Statistics. John Wiley & Sons, Inc., New York, NY, USA. M. Grossman and W.J. Koops. 2003. Modeling Extended Lactation Curves of Dairy Cattle: A Biological Basis for the Multiphasic Approach. J. Dairy Sci. 86:988-998. W. J. Koops. 1986. Multiphasic Growth Curve Analysis. Growth 50:169-177. Jose C. Pinheiro and Douglas M. Bates 2000. Mixed-Effects Models in S and S-Plus. Statistics and Computing. Springer-Verlag New York, New York, NY, USA. W. H. Venables and B. D. Ripley. 2002. Modern Applied Statistics with S. Fourth edition. Statistics and Computing. Springer-Verlag New York, Inc., New York, NY, USA. K. A. Weigel, B. A. Craig, T. R. Bidwell and D. M. Bates. 1992. Comparison of Alternative Diphasic Lactation Curve Models under Bovine Somatotropin Administration. J. Dairy Sci. 75:580-589. Peter B. Mandeville cel: 444 860 3204 tel: 52 444 826 2346-49 ext 532 fax: 52 444 826 2352 P.D. Favor de confirmar la llegada de este correo. Gracias. _________________________________________________________________ Discover the new Windows Vista [[alternative HTML version deleted]]
Although I am not an expert in NLME modelling, looking at your data it
seems to me that there seems to be no growth pattern in it. Try
library(nlme)
library(reshape)
goat<-read.table("clipboard", header=T)
> str(goat)
'data.frame': 18 obs. of 15 variables:
$ NoC : int 52 48 54 60 58 42 44 47 50 56 ...
$ PL1 : int 490 970 1450 620 1370 1200 1200 1000 1300 950 ...
$ PL2 : int 950 1020 1420 1250 1350 920 1560 1100 1150 1250 ...
$ PL3 : int 800 980 1430 1100 1200 1150 1650 1000 1200 1200 ...
$ PL4 : int 900 740 1120 1150 1300 850 1600 870 1030 1280 ...
goat.m <- melt(goat, id="NoC")
levels(goat.m$variable) <- 1:14
goat.m$variable <- as.numeric(as.character(goat.m$variable))
names(goat.m)[2] <- "week"
goat.m$NoC <- ordered(goat.m$NoC)
goat.g <- groupedData(value~week|NoC, data=goat.m)
plot(goat.g)
But maybe I am completely mistaken.
Regards
Petr
r-help-bounces at stat.math.ethz.ch napsal dne 04.09.2007 18:02:09:
>
> Greetings R Help Group,
>
> How does one effect a multiphasic logistic growth model with 4 phases
(e.g. > Koops 1986; Weigel, Craig, Bidwell and Bates 1992; Grossman and Koops
2003) with R.>
> Before writing to the group, the R help archives were searched, the web
was > searched with Google, Venables and Ripley 2002 was consulted, Pinheiro
and > Bates 2000 was consulted, Bates and Watts 2007 was bought and consulted,
ETC. > but to no avail.
>
> I have not written to any other group with respect to this problem.
>
> The following data are offered as an example of the type of problem I am
> dealing with and are average daily goat milk production in ml. for 14
weeks. > NoC is the goat number. PL1 is the production during the first week; PL2
is > the production during the second week, etc.
>
> NoC PL1 PL2 PL3 PL4 PL5 PL6 PL7 PL8 PL9 PL10 PL11 PL12 PL13
PL14> 52 490 950 800 900 850 850 750 610 640 900 980 890 890
910> 48 970 1020 980 740 1050 970 850 790 900 920 1120 1120 1030
1300> 54 1450 1420 1430 1120 1330 1230 1030 1170 1350 1530 1490 1500 1310
910> 60 620 1250 1100 1150 780 930 990 940 760 730 790 1050 840
850> 58 1370 1350 1200 1300 1350 1310 1070 910 1010 1300 1110 1070 990
660> 42 1200 920 1150 850 720 630 630 710 850 810 930 980 820
1570> 44 1200 1560 1650 1600 1450 1600 1160 1010 1440 1450 1530 1500 1550
850> 47 1000 1100 1000 870 760 900 820 865 910 820 930 900 1130
1070> 50 1300 1150 1200 1030 1070 970 860 900 950 1190 1250 1130 800
1400> 56 950 1250 1200 1280 1220 1155 840 1016 1370 1220 1570 1520 1500
1150> 1 870 1250 1160 1270 1200 1410 1110 1008 970 1130 1490 1330 1320
820> 3 1000 1100 1200 1120 1250 980 750 890 1050 1160 1340 1210 1150
760> 4 551 760 550 580 540 620 550 520 470 720 680 790 750
1230> 5 810 1100 820 950 930 830 850 650 810 1070 1120 1300 1040
1320> 6 800 1000 620 850 750 670 660 620 600 610 760 900 758
1070> 7 720 830 1120 1050 820 820 850 810 800 750 780 940 1050
1310> 8 950 1550 1560 1500 1230 1330 1150 1005 1020 1200 1440 1400 1290
1080> 10 660 850 1100 980 1070 1100 870 790 880 950 1000 1210 1050
1220>
> It seems to me that it should be possible to effect the modeling process
with > nlme. Any suggestions and or recommendations would be greatly
appreciated.>
> Peter B.
>
>
>
> Douglas M. Bates and Donald G. Watts. 2007. Nonlinear Regression
Analysis and > Its Applications. Wiley Series in Probability and Statistics. John Wiley
& > Sons, Inc., New York, NY, USA.
>
> M. Grossman and W.J. Koops. 2003. Modeling Extended Lactation Curves of
Dairy > Cattle: A Biological Basis for the Multiphasic Approach. J. Dairy Sci.
86:988-998.>
> W. J. Koops. 1986. Multiphasic Growth Curve Analysis. Growth 50:169-177.
>
> Jose C. Pinheiro and Douglas M. Bates 2000. Mixed-Effects Models in S
and S-> Plus. Statistics and Computing. Springer-Verlag New York, New York, NY,
USA.>
> W. H. Venables and B. D. Ripley. 2002. Modern Applied Statistics with S.
> Fourth edition. Statistics and Computing. Springer-Verlag New York,
Inc., New > York, NY, USA.
>
> K. A. Weigel, B. A. Craig, T. R. Bidwell and D. M. Bates. 1992.
Comparison of > Alternative Diphasic Lactation Curve Models under Bovine Somatotropin
> Administration. J. Dairy Sci. 75:580-589.
>
>
>
>
>
>
> Peter B. Mandeville cel: 444 860 3204 tel: 52 444 826 2346-49 ext
532 > fax: 52 444 826 2352 P.D. Favor de confirmar la llegada de este correo.
Gracias.> _________________________________________________________________
> Discover the new Windows Vista
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html> and provide commented, minimal, self-contained, reproducible code.
Dear Petr,
I tried replying directly to you but my communication was rejected as spam.
On second thought, it is possible that it is.
Thank you very much for taking the time and interest to answer my enquiry to
the R Help Group.
My interest isn’t so much in the particular data set as in how to conduct
such an analysis in R. The URL of the article where Bates was a coauthor
which describes the technique is
<http://jds.fass.org/cgi/content/abstract/75/2/580>
http://jds.fass.org/cgi/content/abstract/75/2/580
I enclose the IV Rabbit weight (age (a) in days, weight (w) in grams) data
set from Koops 1986:171 (citation given in the original communication).
a w
0 35
10 103
20 181
30 300
40 395
40 479
60 561
70 641
80 720
90 810
100 892
110 958
120 1013
130 1072
140 1118
150 1168
160 1195
170 1208
The published results are (Koops 1986:173)
Table 2: Constant estimates for the multiphasic growth curve
Data set a1 b1 c1 a1 b2 c2
a3 b3 c3
Rabbit 198 .048 24.8 352 .025 82.7 62
.051 140.2
Constants ai is in grams and ci in days. Constants bi in 1/(age-units).
Table 3. Information about the fit in a stepwise procedure
Data set number of error residual R-square
Durbin-Watson
phases(n) df variance
statistic
Rabbit 1 15 1268
.9928 .35
2 12 97
.9996 1.39
3 9 40
.9999 2.78
4 6 47
.9999 3.19
My question is if it is possible to effect a multiphasic analysis in R and
if so how?
Thank you very much,
Peter B.
Laboratorio de Informática, Facultad de Medicina, UASLP, SLP, SLP, MEX
Departamento de Epidemiología Clínica, Facultad de Medicina, UASLP, SLP,
SLP, MEX
Peter B. Mandeville
cel: 444 860 3204
tel: 52 444 826 2346-49 ext 532
fax: 52 444 826 2350
P.D. Favor de confirmar la llegada de este correo. Gracias.
[[alternative HTML version deleted]]