[R] xyplot w/ panel.lmline solution
Dear Sebastian Luque (and All R Users) With the following code I managed to plot different characters and regression lines for panels 2 (Day of year 101) and 4 (Days of year 151, 157 and 172): xyplot(log(no.larvae)~age.cls|factor(day),data=mortal, layout=c(7,1),aspect=5/3, xlab=Age class (d),ylab=Ln(Abundance), ylim=c(-2.5,6.5),xlim=c(0,30), panel = function(x, y,panel.number) { if(panel.number==2){ panel.xyplot(x, y,col=1) panel.xyplot(x[6:11],y[6:11],pch=16,col=1) panel.lmline(x[6:11],y[6:11])} else{ panel.xyplot(x, y,col=1) panel.xyplot(x[2:11],y[2:11],pch=16,col=1) panel.lmline(x[2:11],y[2:11])} if(panel.number4){ panel.xyplot(x, y,col=1) panel.xyplot(x[14:20],y[14:20],pch=16,col=2) panel.lmline(x[14:20],y[14:20],lty=2)} }) The resulting plot is herein (attached MortalityRates.pdf file). Thanks to you all, Eduardo Esteves MortalityRates.pdf Description: Adobe PDF document __ R-help@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
[R] help w/ xyplot and panel.lmline
Dear All, I'm confortable with xyplot(...) and panel.lmline(...) statements (at least I thought I did :). I've used the following code to plot the decline in log-abundance of fish larvae (no.larvae) with age (age.cls, 4 to 27 days-old) for specific dates of sampling (day, 9 dates). I further plotted data with different colors and regression lines for ages 5-14 d and 17-23 d in a 7-by-1 layout. xyplot(log(no.larvae)~age.cls|factor(day),data=mortal, layout=c(7,1),aspect=5/3, xlab=Age class (d),ylab=Ln(Abundance),ylim=c(-2.5,6.5),xlim=c(0,30), panel = function(x, y) { panel.xyplot(x, y, col=1) panel.xyplot(x[5:14],y[5:14],pch=16,col=1) panel.lmline(x[5:14],y[5:14]) panel.xyplot(x[17:23],y[17:23],pch=16,col=2) panel.lmline(x[17:23],y[17:23],lty=2) }) Is it possible to change the plotted characters and regression lines for two of the panels (corresponding to dates 101 and 172). For these dates I intend to use data only for ages 9-14 d instead of 5-14 d as for the remaining. Thanks in advance, Eduardo Esteves __ R-help@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
[R] Help with multicomarisons after ANCOVA
Dear All, I' ve used ANCOVA (through lm(Y~factor*x)) to study the influence/differences between levels of factor upon the Y-x relationship. I found that both intercepts and slopes differ among levels of the factor. I understand from the results of summary(lm(Y~factor*x)) which of the intercepts/slopes per level differ from the reference level. I later used lm(Y~factor/(1+x)-1) to obtain explicit coefficients per level. My question is: Can I test for differences among all intercepts and /or slopes per levels using simtest (from package multicomp) as summary (simtest(Y~factor+x,type=tukey))? The outputs seem to contradict the results from the explicit coefficients approach! Thanks in advance, Eduardo __ R-help@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
[R] help with vector operation
Dear All, I'm trying to calculate the following: g.rate-(SL-int)/NO where SL and NO are individual length and counts for the same subject, and the int value is a day-specific value i.e. for each DAY=88, 101..., 172 I'd like to use a different values of int=9.32, 8.43, ..., 9.81. All variables are compiled in a larger data.frame of the form: CODE SLNO DAY (...) 123 12.5 14 88 124 11.4 13 88 125 10.8 11 101 (...) 234 16.9 19 172 (...) I've been trying apply() but with little (NONE) success! Thanks, Eduardo Esteves __ R-help@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
[R] Help w/ matrix calc
Dear All, Help is needed! I have a matrix with frequencies of fish larvae per length class (var. sl) and age-group (var. median.no) obtained with k-table(cut(sl,(5:22)),median.no) k[2:5,1:5] #to ilustrate k 4 5 6 7 (6,7] 3 1 0 0 (7,8] 3 0 1 0 (8,9] 3 4 3 5 (9,10] 3 15 7 13 from this matrix I would like to obtain the mean age per length class i.e. vector of line means. How can I do this? If, instead of the means I wanted the variances, how could I do it? Thanks in advance, Eduardo Esteves __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] lme question
Dear All, I need help both with analytical and computational aspects. My problem: I have counted and measured increments on otoliths (inner-ear bony structures) of fish larvae collected at different times of day. This was repeated three times for each fish larvae (in random order and no particular dates). Then, I calculated the ratio of marginal-to-previous increment for each fish and reading. I wish to study if differences exist in ratios between hours of day (accounting for differences in readings within each fish) - in data.frame format... code hour read.no ratio 33001 5 1 0.56 33003 7 1 0.93 37123 19 1 0.23 ... 33001 5 2 0.62 33003 7 2 0.87 37123 19 2 0.22 ... 37123 19 3 0.26 ... My solution: Use function lme, with hr-factor(hour)... hr.lme-lme(sqrt(ratio)~hr,random=~1|code/read.no) anova(hr.lme) ; summary (hr.lme) Does this commands correspond to the initial problem statement? Thanks in advance. Regards, Eduardo Esteves __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] model II regression
Dear All, A friend asked me for help with a regression problem dealing with y and x, both random variables. The x and y are respectively the observed and estimated biomass values of a green algae derived from an ecological model. We first considered it to be a model II regression and derived the intercept and slope using Bartlett's method, reduced major axis (geometric mean regression) and major axis method (as implemented by Legendre). Then the significance and percentage of variance explained by the models was calculated (by hand) through ANOVA or Permutation test/Correlation (for later mehod). Our questions are: using a special approach to regression (model II) and then an ANOVA makes sense? I favour Legendre's approach but cannot justiy why! Is there any code/package available dealing with this problem? Thanks, Eduardo __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] factorial experiments and repeated measures
Dear All, I'm new to the list and relatively new to R (but not so to S-Plus) and I´ve been asked to help study some experimental data. In a factorial experiment (5 factors at 2 levels) the response-variable was measured at various times (0, 2, 5, 7, ..., 30 days after start). There's just one run for each factor combination! I intend to use the nlme library of Pinheiro and Bates in a longitudinal data situation but I have some difficulties with the statistical scenario of it all! The factors are not nested and the number of repeated measures is unbalanced (in some cases they were abandoned for some reason before the end of the experiment). The main objectives are to identify the factor(s) that influence the response and to model it. How should I dissecate this? If in need of more info, just ask! Thanks. Best regards, Eduardo Esteves __ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help