gexp: Generator of Experiments
In gexp: Generator of Experiments
gexp R Documentation
Generator of Experiments
Description
The package provides computational resources useful in planning and modeling of structured experiments in the R environment.
The generic function S3 gexp was created to enable plan, create and model structured experiments, that is, under a design experimental. In the modeling it is possible to simulate results of experiments with possibility of user informing the effects and the random error(s). The designs are: Completely Randomized Design (CRD), Randomized complete block design (RCBD) and Latin Squares Design (LSD). The types of experiments are: Factorial Experimentation (FE) and Split-plot experiment (SPE).
The experiments can be generated with one or more variable response, in the latter case, it may be important for a structure covariance between them. It is also possible to plan experiments with of graphic parts for use in planning from figures or pictures of the experimental area.
Usage
gexp(x, ...)
## Default S3 method:
gexp(
x = NULL,
mu = 26,
err = NULL,
errp = NULL,
r = 5L,
fl = NULL,
blkl = NULL,
rowl = NULL,
coll = NULL,
fe = NULL,
inte = NULL,
blke = NULL,
rowe = NULL,
cole = NULL,
contrasts = NULL,
type = c('SIMPLE','FE','SPE'),
design = c('CRD','RCBD','LSD'),
round = 2L,
...)
## S3 method for class 'simple_crd'
gexp(
x, ...)
## S3 method for class 'simple_rcbd'
gexp(
x, ...)
## S3 method for class 'simple_lsd'
gexp(
x, ...)
## S3 method for class 'fe_crd'
gexp(
x, ...)
## S3 method for class 'fe_rcbd'
gexp(
x, ...)
## S3 method for class 'fe_lsd'
gexp(
x, ...)
## S3 method for class 'spe_crd'
gexp(
x, ...)
## S3 method for class 'spe_rcbd'
gexp(
x, ...)
## S3 method for class 'spe_lsd'
gexp(
x, ...)
Arguments
x
An object of gexp class.
mu
Is a numeric scalar, or a vector to Multivarite Data (MD), that represent the mean of each factor. Required NULL if the effect of at least one factor of the experiment is quantitative.
err
It is a vector, or matrix for MD, that represents the experimental error. The default value is NULL, that is, for each response variable a normal error is added with mean 0 and variance 1 generated by rmvnorm(sigma = diag(length(mu))).
errp
It is a vector, or a matrix for MD, of the error associated with the plots if type is equal to SPE (Split-Plot Experiments). The default value is NULL, that is, for each response variable a normal error is added with mean 0 and variance 1 generated by rmvnorm(sigma = diag(length(mu))).
r
It is a scalar of the number of repetitions of the experiment.
fl
List of a vector of characters, or a matrix (MD). It's a list of factor names.
blkl
List of a vector of characters, or an array for MD, of block names.
rowl
List a vector of characters, or an array for MD, of the line names in case type is equal to LSD (Latin Square Design).
coll
List of a vector of characters, or an array for MD, of the column names in case the type is equal to LSD (Latin Square Design).
fe
It is a numerical vector, or a matrix (MD). It's a list of the effect of a factor.
inte
It is a numerical vector, or a matrix for MD, of the effects of the interaction.
blke
It is a numerical vector, or a matrix for MD, of the effects of the blocks.
rowe
It is a numerical vector, or an array for MD, of the effects of the lines in case the type is equal to LSD (Latin Square Design).
cole
Is a numeric vector, or a matrix for MD, of the effects of the columns in case the type is equal to LSD (Latin Square Design).
contrasts
A list, whose entries are values (numeric matrices or character strings naming functions) to be used as replacement values for the contrasts function and whose names are the names of the columns. See "contrasts.arg" argument of the model.matrix function to more details. Required if the effect of at least one factor of the experiment is quantitative.
type
It is a vector of strings that contains the type of experiment to be used: Simple (SIMPLE), Factorial Experiment (FE) and Split-plot Experiment (SPE). SIMPLE is the default.
design
It is a vector of strings that contains the type of design to be used: Completely Randomized Design (CRD), Randomized Complete Block Design (RCBD), Latin Squares Design (LSD). CRD is the default.
round
This is a numeric scalar for rounding of the response variable.
...
Further arguments (required by generic).
Value
The method gexp returns the list of class gexp with the slots:
X
It is the incidence matrix of the design.
Z
It is the incidence matrix of the error of the main parcel in the case of type equal to SPLIT.
Y
It is a vector, or a matrix for MD, with the values of the random variable(s).
dfm
It is a data.frame with all experiment information: treatments, repetitions, and the random response variable.
References
Ferreira, Daniel Furtado. 2008. Estatistica Multivariada. 1 ed.
Lavras: Ed. UFLA.
Aquino, Luiz Henrique. Tecnica Experimental com Animais I. Apostila da
disciplina “Tecnica Experimental com Animais” da Universidade Federal de
Lavras, 1992.
Rencher, Alvin C. and Schaalje, Bruce G. 2007. Linear Models in
Statistics, second edition. Hoboken: John Wiley and Sons.
Naes, T.; Aastveit, A.H.; Sahni, N.S. 2007. "Analysis of split-plot designs:
An Overview and Comparison of Methods". Qual. Reliab. Engng. Int. 23,
801-820.
See Also
plot.gexp.simple_crd
Examples
#++++++++++++++++++++++++++++#
# UNIVARIATE APPROACH #
#++++++++++++++++++++++++++++#
#!____________________________
#! Qualitative Factor(s) (QL)
#!____________________________
#! Completely Randomized Design (CRD)
#! 1 factor - CRD - QLF
# Nonsense(experimental error = 0)
# Yi = mu + fe + e
r <- 2 # (repet. number)
fln <- 3 # (factor levels number)
crd00 <- gexp(mu = 0,
r = r,
fe = list(f1 = c(1, 2, 3)),
err = matrix(0,
nrow = r*fln),
round = 0)
crd00$X
print(crd00)
summary(crd00)
str(crd00)
#! 1 factor - CRD - QL
# Nonsense(error is 0)
# Yi = mu + fe + e
r <- 3 # (repet. number)
fln <- 5 # (factor levels number)
crd01 <- gexp(mu = 1,
r = r,
fe = list(f1 = c(0, 2, 4, 6, 8)),
err = matrix(0,
nrow = r*fln),
round = 2)
summary(crd01)
#! 1 factor - CRD - QL
# Default error: rmvnorm(sigma = diag(ncol(as.matrix([[fe]]))))
crd_1f <- gexp(mu = 1,
r = 3,
fe = list(f1 = c(1, 1, 5, 1, 1)),
fl = list(Treat = LETTERS[1:5]),
round = 2)
crd_1f$X
summary(crd_1f)
#! Binomial error - CRD - QL
e_binom <- as.matrix(rbinom(n = 15,
size = 5,
prob = 0.1))
crd_bin <- gexp(mu = 20,
err = e_binom,
r = 5,
fe = list(f1 = c(1, 4, 1)))
summary(crd_bin)
mod <- aov(Y1 ~ X1,
data = crd_bin$dfm)
shapiro.test(mod$res)
#! Factorial Experiment (FE) - CRD - QL
fe_crd00 <- gexp(mu = 0,
r = 2,
fe = list(f1 = c(1, 1, 5),
f2 = c(1, 1),
f3 = c(2, 2, 1)),
fl = list(A = paste('a',
1:3,
sep = ''),
B = paste('b',
1:2,
sep = ''),
C = paste('c',
1:3,
sep = '')),
round = 0,
type = 'FE')
fe_crd00$X
summary(fe_crd00)
#! Factorial Experiment (FE) - With interaction - CRD - QL
fe_crd01 <- gexp(mu = 30,
fe = list(f1 = c(1, 1, 3),
f2 = c(1, 1)),
fl = list(A = paste('a',
1:3,
sep = ''),
B = paste('b',
1:2,
sep = '')),
inte = c(3, 1, 1, 1, 1, 5), # (3*2)
round = 1,
type = 'FE')
summary(fe_crd01)
#! Split-plot Experiment (SPE) - CRD - QL
split_crd <- gexp(mu = 30,
fe = list(f1 = c(1, 1),
f2 = c(2, 3)),
fl = list(P = paste('p',
1:2,
sep = ''),
SP = paste('sp',
1:2,
sep = '')),
inte = c(1, 15, 1, 1), # (2*2)
round = 1,
type = 'SPE',
design = 'CRD')
split_crd$X
split_crd$Z
summary(split_crd)
split_crd01 <- gexp(mu = 30,
r = 3,
fe = list(f1 = c(1, 1),
f2 = c(2, 3),
f3 = c(1, 1, 1)),
fl = list(P = paste('p',
1:2,
sep = ''),
A = paste('a',
1:2,
sep = ''),
B = paste('b',
1:3,
sep = '')),
round = 1,
type = 'SPE',
design = 'CRD')
split_crd01$X
split_crd01$Z
summary(split_crd01)
#! Randomized Complete Block Design (RCBD) - QL
# 1 factor, 3 blocks
rcbd <- gexp(mu = 0,
r = 2,
fe = list(f1 = c(5, 1, 1)),
fl = list(TR = LETTERS[1:3]),
blke = c(1, 2, 3),
blkl = list(BLK = paste('B',
1:3,
sep = '')),
round = 1,
design = 'RCBD')
rcbd$X
summary(rcbd)
#! Factorial Experiment (FE) - RCBD - QL
fe_rcbd <- gexp(mu = 30,
r = 2,
fe = list(f1 = c(1, 1, 1),
f2 = c(2, 3)),
blke = c(1, 3),
inte = c(1, 15, 1, 1, 5, 1), # (3*2)
round = 1,
type = 'FE',
design = 'RCBD')
summary(fe_rcbd)
#! Multivariated - RCBD - QL
rcbd_m <- gexp(mu = c(0, 2),
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
1, 2,
1, 1),
ncol = 2,
byrow = TRUE),
round = 1,
design = 'RCBD')
summary(rcbd_m)
#! Split-plot Experiment (SPE) - RCBD - QL
split_rcbd <- gexp(mu = 30,
r = 2,
fe = list(f1 = c(1, 1),
f2 = c(2, 3),
f3 = c(1, 1, 1)),
fl = list(P = paste('p',
1:2,
sep = ''),
B = paste('b',
1:2,
sep = ''),
C = paste('c',
1:3,
sep = '')),
blke = c(1, 2),
blkl = list(BLK = paste('B',
1:2,
sep = '')),
round = 1,
type = 'SPE',
design = 'RCBD')
split_rcbd$Z
summary(split_rcbd)
#! Latin Square Design (LSD) - QL
#!. Warning!!!! r = 5 by default
lsd00 <- gexp(design = 'LSD')
#Set r = 1 to omiting warning
lsd01 <- gexp(mu = 30,
r = 1,
fe = list(f1 = c(1, 1, 10)),
rowe = c(1, 1, 1),
cole = c(1, 1, 1),
rowl = list(Row = paste('r',
1:3,
sep = '')),
coll = list(Col = paste('c',
1:3,
sep = '')),
round = 0,
design = 'LSD')
summary(lsd01)
#! Factorial Experiment (FE) - LSD - QL
fe_lsd <- gexp(mu = 30,
r = 1,
fe = list(f1 = c(1, 1),
f2 = c(2, 3)),
rowe = c(1, 3, 2, 1),
cole = c(2, 2, 1, 1),
rowl = list(Row = paste('r',
1:4,
sep = '')),
coll = list(Col = paste('c',
1:4,
sep = '')),
inte = c(1, 15, 1, 1), # (2*2)
round = 1,
type = 'FE',
design = 'LSD')
summary(fe_lsd)
#! Split-plot Experiment (SPE) - LSD - QL
split_lsd <- gexp(mu = 30,
r = 1,
fe = list(f1 = c(1, 1, 2),
f2 = c(2, 3, 1)),
fl = list(P = paste('p',
1:3,
sep = ''),
SP = paste('sp',
1:3,
sep = '')),
inte = c(1, 15, 1, 1, 1, 1, 1, 1, 1), # (3*3)
rowe = c(1, 1, 1),
cole = c(1, 1, 1),
rowl = list(Row = paste('r',
1:3,
sep = '')),
coll = list(Col = paste('c',
1:3,
sep = '')),
round = 1,
type = 'SPE',
design = 'LSD')
summary(split_lsd)
#!_____________________________
#! Quantitative Factor(s) (QT)
#!_____________________________
#! CRD - Orthogonal polynomials
# Linear effect
# Nonsense(error is 0)
# Default contrasts: Orthogonal contrasts
r <- 4 # (repet. number)
fln <- 5 # (factor levels number)
level <- c(0, 10, 20, 30, 40)
crd_lo <- gexp(mu = 1, #in this case, mu=beta0 (intercept)
r = r,
fe = list(f1 = c(2, 0, 0, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
err = matrix(0,
nrow = r*fln),
round = 2)
crd_lo$X
summary(crd_lo)
plot(Y1 ~ Dose,
crd_lo$dfm)
# Quadratic effect
crd_qo <- gexp(mu = 2,
r = r,
fe = list(f1 = c(0, 3, 0, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
err = matrix(0,
nrow = r*fln))
summary(crd_qo)
plot(Y1 ~ Dose,
crd_qo$dfm)
# Cubic effect
crd_co <- gexp(mu = 2,
r = r,
fe = list(f1 = c(1, 1, 3, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
err = matrix(0,
nrow = r*fln))
summary(crd_co)
plot(Y1 ~ Dose,
crd_co$dfm)
# Not orthogonal polynomials
# Linear
cont_crd <- matrix(c(level,
level^2,
level^3,
level^4),
ncol = 4)
crd_l <- gexp(mu = 2,
r = 2,
fe = list(f1 = c(10, 0, 0, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
contrasts = list(Dose = cont_crd))
crd_l$X
summary(crd_l)
plot(Y1 ~ Dose,
crd_l$dfm)
reg <- lm(Y1 ~ Dose + I(Dose^2) + I(Dose^3) + I(Dose^4),
data = crd_l$dfm)
summary(reg)
# Linear and quadratic
level1 <- seq(0,30,by = 10)
cont_crd1 <- matrix(c(level1,
level1^2,
level1^3),
ncol = 3)
level2 <- 1:4
cont_crd2 <- matrix(c(level2,
level2^2,
level2^3),
ncol = 3)
crd_lq <- gexp(mu = 1,
r = 2,
fe = list(f1 = c(10, 0, 0), #b1 #b2 #b3
f2 = c(1, 8, 0)),
fl = list(P = level1,
N = level2),
contrasts = list(N = cont_crd2,
P = cont_crd1))
crd_lq$X
summary(crd_lq)
with(crd_lq$dfm,
plot(Y1 ~ P))
with(crd_lq$dfm,
plot(Y1 ~ N))
# Multivariated
crd_m <- gexp(mu = c(2, 10),
r = 4,
fe = list(f1 = matrix(c(10, 0, #L Q
0, 10,
0, 0),
ncol = 2,
byrow = TRUE)),
fl = list(Dose = level1),
contrasts = list(Dose = cont_crd1))
with(crd_m$dfm,
plot(Y1 ~ Dose))
with(crd_m$dfm,
plot(Y2 ~ Dose))
# RCBD - Orthogonal polynomios
level3 <- c(0, 2, 4, 6)
rcbd <- gexp(mu = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
blke = c(1, 2, 3),
r = 2,
fl = list(Dose = level3),
blkl = list(Blk = c('B1', 'B2', 'B3')),
design = 'RCBD')
rcbd$X
summary(rcbd)
# Not orthogonal
cont_crd3 <- matrix(c(level3, level3^2, level3^3),
ncol = 3)
rcbd_01 <- gexp(mu = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
blke = c(1, 2, 3),
r = 2,
fl = list(Dose = level3),
blkl = list(Blk = c('B1', 'B2', 'B3')),
contrasts = list(Dose = cont_crd3),
design = 'RCBD')
rcbd_01$X
summary(rcbd_01)
# Orthogonal polynomios - LSD
lsd <- gexp(mu = 1,
r = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
rowe = rep(1, 4),
cole = rep(1, 4),
fl = list(Dose = level1),
design = 'LSD')
lsd$X
summary(lsd)
lsd_01 <- gexp(mu = 1,
r = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
rowe = rep(1, 4),
cole = rep(1, 4),
rowl = list(row = paste('r',
1:4,
sep = '')),
fl = list(Dose = level1),
design = 'LSD')
lsd_01$X
summary(lsd_01)
# Not orthogonal
lsd_02 <- gexp(mu = 1,
r = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
rowe = rep(1, 4),
cole = rep(1, 4),
fl = list(Dose = level3),
contrasts = list(Dose = cont_crd3),
design = 'LSD')
lsd_02$X
str(lsd_02)
#!__________________________________________________________________________
#! Hibrid: qualitative and quantitative factors in the same experiment - HB
#!__________________________________________________________________________
#! CRD - HB
r <- 2 # (repet. number)
fl1 <- 4# (first factor levels number)
fl2 <- 3# (second factor levels number)
crd_hb <- gexp(mu = 1, #in this case, mu=beta0 (intercept)
r = r,
fe = list(f1 = c(2, 0, 0), #b1 #b2 #b3
f2 = c(1, 1, 3)),
fl = list(Dose = seq(0,30,
by = 10),
Trat = LETTERS[1:3]),
err = matrix(0,
nrow = r*fl1*fl2),
round = 2)
crd_hb$X
summary(crd_hb)
#Only one contrasts!
crd_hb2 <- gexp(mu = 1, #in this case, mu=beta0 (intercept)
r = r,
fe = list(f1 = c(2, 0, 0), #b1 #b2 #b3
f2 = c(1, 1, 3)),
fl = list(Dose = level1,
Trat = LETTERS[1:3]),
err = matrix(0,
nrow = r*fl1*fl2),
contrasts = list(Dose = cont_crd1),
round = 2)
crd_hb2$X
summary(crd_hb)
#! RCBD - HB
r <- 2
blke <- c(1, 2)
level <- c(0, 10, 20, 30)
(error <- matrix(rep(0,
4^1*3^1*r*length(blke)),
ncol=1))
rcbd_hb <- gexp(mu = 2,
err = error,
r = r,
fe = list(f1 = c(0, 1, 0), # Qualitative
f2 = c(1, 0, 0)), # Quantitative linear
fl = list(Var = LETTERS[1:3],
Dose = level),
blke = blke,
blkl = list(Blk = c('B1', 'B2')),
design = 'RCBD')
rcbd_hb$X
summary(rcbd_hb)
str(rcbd_hb)
#! LSD - QT
set.seed(3)
lsd <- gexp(mu = 100,
r = 1,
fe = list(f1 = c(10, # b1
20, # b2
0, # b3
0)), # b4
fl = list(tra = seq(0,
40,
by = 10)),
rowe = c(1, 2, 3, 4, 5),
rowl = list(row = paste('r',
1:5,
sep = '')),
cole = c(5, 4, 3, 2, 1),
coll = list(col = paste('c',
1:5,
sep = '')),
design = 'LSD')
summary(lsd)
plot(Y1 ~ tra, lsd$dfm)
#! FE - LSD - QT
fe_lsd <- gexp(mu = 10,
fe = list(f1 = c(2, 3),
f2 = c(5, # b1*
0, # b2
0, # b3
0)), # b4
rowe = rep(1, 10),
cole = rep(1, 10),
fl = list(var = paste('v',
1:2,
sep = ''),
tra = seq(0,
40,
by = 10)),
coll = list(col = paste('c',
1:10,
sep = '')),
rowl = list(row = paste('r',
1:10,
sep = '')),
type = 'FE',
design = 'LSD')
fe_lsd$X
str(fe_lsd)
summary(fe_lsd)
plot(Y1 ~ tra,
fe_lsd$dfm)
#! SPE - QL - QT
spe_lsd <- gexp(mu = 100,
r = 1,
fe = list(f1 = c(2, 3, 1),
f2 = c(1, # b1
5, # b2*
1)), # b3
fl = list(p = paste('p',
1:3,
sep = ''),
sp = seq(0,
30,
by = 10)),
rowe = c(1, 2, 3),
cole = c(3, 2, 1),
rowl = list(row = paste('r',
1:3,
sep = '')),
coll = list(col = paste('c',
1:3,
sep = '')),
round = 1,
type = 'SPE',
design = 'LSD')
summary(spe_lsd)
plot(spe_lsd)
#++++++++++++++++++++++++++++#
# MULTIVARIATE APPROACH #
#++++++++++++++++++++++++++++#
#! CRD - QL
# Error = 0 - Nonsense (you can easily undertand the effects)
r <- 2 # (repet. number)
fln <- 3 # (factor levels number)
crd_m01 <- gexp(mu = c(0,10),
r = r,
fe = list(f1 = matrix(c(1, 0, #Y1 Y2
2, 1,
3, 3),
ncol = 2,
byrow = TRUE)),
err = mvtnorm::rmvnorm(n = fln * r,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
round = 0)
summary(crd_m01)
#! FE - CRD - QL
r <- 2
crd_m_fe01 <- gexp(mu = c(0, 0),
r = r,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * r,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
type = 'FE',
round = 1)
summary(crd_m_fe01)
#! FE - CRD - QL
# Using default error
set.seed(30)
crd_m_fe02 <- gexp(mu = c(0, 2),
r = 3,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
type = 'FE',
round = 1)
summary(crd_m_fe02)
#! SPE - CRD - QL
# Using default error
crd_m_spe01 <- gexp(mu = c(0, 2),
r = 3,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
type = 'SPE',
round = 1)
summary(crd_m_spe01)
#! RCBD - QL
r <- 2 # (repet. number)
fln <- 3 # (factor levels number)
bln <- 3 # (block levels number)
rcbd_m01 <- gexp(mu = c(0,10),
r = r,
fe = list(f1 = matrix(c(1, 0, #Y1 Y2
2, 1,
3, 3),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
err = mvtnorm::rmvnorm(n = fln * r * bln,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
design = 'RCBD',
round = 0)
summary(rcbd_m01)
#! FE - RCBD - QL
rcbd_m_fe01 <- gexp(mu = c(0, 0),
r = r,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * r * bln,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'FE',
design = 'RCBD',
round = 1)
summary(rcbd_m_fe01)
#! SPE - RCBD - QL
rcbd_m_spe01 <- gexp(mu = c(0, 2),
r = 2,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE),
f3 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'SPE',
design = 'RCBD',
round = 1)
summary(rcbd_m_spe01)
#! LSD - QL
lsd_m01 <- gexp(mu = c(0,10),
r = 1,
fe = list(f1 = matrix(c(1, 0,
2, 1,
3, 3),
ncol = 2,
byrow = TRUE)),
rowe = matrix(rep(1, 6),
ncol = 2),
cole = matrix(rep(1, 6),
ncol = 2),
err = mvtnorm::rmvnorm(n = 3^2,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
design = 'LSD',
round = 0)
summary(lsd_m01)
#! LSD/FE - QL
lsd_m_fe01 <- gexp(mu = c(0, 0),
r = 1,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * 6,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
#Y1 Y2
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
#Y1 Y2
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
rowe = matrix(rep(1, 12),
ncol = 2),
cole = matrix(rep(1, 12),
ncol = 2),
type = 'FE',
design = 'LSD',
round = 1)
summary(lsd_m_fe01)
#! SPE - LSD - QL
# Using default error
lsd_m_spe01 <- gexp(mu = c(0, 2),
r = 1,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
rowe = matrix(rep(1, 6),
ncol = 2),
cole = matrix(rep(1, 6),
ncol = 2),
type = 'SPE',
design = 'LSD',
round = 1)
summary(lsd_m_spe01)
#! FE - RCBD - QL
r = 1
bln = 3
fe_rcbd_m <- gexp(mu = c(0, 0),
r = 1,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * r * bln,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'FE',
design = 'RCBD')
str(fe_rcbd_m)
summary(fe_rcbd_m)
#! SPE - RCBD - QL
spe_rcbd_m <- gexp(mu = c(0, 2),
r = 3,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE),
f3 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'SPE',
design = 'RCBD')
str(spe_rcbd_m)
summary(spe_rcbd_m)
gexp documentation built on July 9, 2023, 5:42 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| gexp | R Documentation |
Generator of Experiments
Description
The package provides computational resources useful in planning and modeling of structured experiments in the R environment.
The generic function S3 gexp was created to enable plan, create and model structured experiments, that is, under a design experimental. In the modeling it is possible to simulate results of experiments with possibility of user informing the effects and the random error(s). The designs are: Completely Randomized Design (CRD), Randomized complete block design (RCBD) and Latin Squares Design (LSD). The types of experiments are: Factorial Experimentation (FE) and Split-plot experiment (SPE).
The experiments can be generated with one or more variable response, in the latter case, it may be important for a structure covariance between them. It is also possible to plan experiments with of graphic parts for use in planning from figures or pictures of the experimental area.
Usage
gexp(x, ...)
## Default S3 method:
gexp(
x = NULL,
mu = 26,
err = NULL,
errp = NULL,
r = 5L,
fl = NULL,
blkl = NULL,
rowl = NULL,
coll = NULL,
fe = NULL,
inte = NULL,
blke = NULL,
rowe = NULL,
cole = NULL,
contrasts = NULL,
type = c('SIMPLE','FE','SPE'),
design = c('CRD','RCBD','LSD'),
round = 2L,
...)
## S3 method for class 'simple_crd'
gexp(
x, ...)
## S3 method for class 'simple_rcbd'
gexp(
x, ...)
## S3 method for class 'simple_lsd'
gexp(
x, ...)
## S3 method for class 'fe_crd'
gexp(
x, ...)
## S3 method for class 'fe_rcbd'
gexp(
x, ...)
## S3 method for class 'fe_lsd'
gexp(
x, ...)
## S3 method for class 'spe_crd'
gexp(
x, ...)
## S3 method for class 'spe_rcbd'
gexp(
x, ...)
## S3 method for class 'spe_lsd'
gexp(
x, ...)
Arguments
x |
An object of gexp class. |
mu |
Is a |
err |
It is a vector, or |
errp |
It is a vector, or a |
r |
It is a scalar of the number of repetitions of the experiment. |
fl |
List of a vector of characters, or a |
blkl |
List of a vector of characters, or an array for MD, of block names. |
rowl |
List a vector of characters, or an |
coll |
List of a vector of characters, or an |
fe |
It is a numerical vector, or a |
inte |
It is a numerical vector, or a |
blke |
It is a numerical vector, or a |
rowe |
It is a numerical vector, or an |
cole |
Is a |
contrasts |
A |
type |
It is a vector of strings that contains the type of experiment to be used: Simple (SIMPLE), Factorial Experiment (FE) and Split-plot Experiment (SPE). SIMPLE is the default. |
design |
It is a vector of strings that contains the type of design to be used: Completely Randomized Design (CRD), Randomized Complete Block Design (RCBD), Latin Squares Design (LSD). CRD is the default. |
round |
This is a numeric scalar for rounding of the response variable. |
... |
Further arguments (required by generic). |
Value
The method gexp returns the list of class gexp with the slots:
X |
It is the incidence |
Z |
It is the incidence |
Y |
It is a vector, or a |
dfm |
It is a |
References
Ferreira, Daniel Furtado. 2008. Estatistica Multivariada. 1 ed. Lavras: Ed. UFLA.
Aquino, Luiz Henrique. Tecnica Experimental com Animais I. Apostila da disciplina “Tecnica Experimental com Animais” da Universidade Federal de Lavras, 1992.
Rencher, Alvin C. and Schaalje, Bruce G. 2007. Linear Models in Statistics, second edition. Hoboken: John Wiley and Sons.
Naes, T.; Aastveit, A.H.; Sahni, N.S. 2007. "Analysis of split-plot designs: An Overview and Comparison of Methods". Qual. Reliab. Engng. Int. 23, 801-820.
See Also
plot.gexp.simple_crd
Examples
#++++++++++++++++++++++++++++#
# UNIVARIATE APPROACH #
#++++++++++++++++++++++++++++#
#!____________________________
#! Qualitative Factor(s) (QL)
#!____________________________
#! Completely Randomized Design (CRD)
#! 1 factor - CRD - QLF
# Nonsense(experimental error = 0)
# Yi = mu + fe + e
r <- 2 # (repet. number)
fln <- 3 # (factor levels number)
crd00 <- gexp(mu = 0,
r = r,
fe = list(f1 = c(1, 2, 3)),
err = matrix(0,
nrow = r*fln),
round = 0)
crd00$X
print(crd00)
summary(crd00)
str(crd00)
#! 1 factor - CRD - QL
# Nonsense(error is 0)
# Yi = mu + fe + e
r <- 3 # (repet. number)
fln <- 5 # (factor levels number)
crd01 <- gexp(mu = 1,
r = r,
fe = list(f1 = c(0, 2, 4, 6, 8)),
err = matrix(0,
nrow = r*fln),
round = 2)
summary(crd01)
#! 1 factor - CRD - QL
# Default error: rmvnorm(sigma = diag(ncol(as.matrix([[fe]]))))
crd_1f <- gexp(mu = 1,
r = 3,
fe = list(f1 = c(1, 1, 5, 1, 1)),
fl = list(Treat = LETTERS[1:5]),
round = 2)
crd_1f$X
summary(crd_1f)
#! Binomial error - CRD - QL
e_binom <- as.matrix(rbinom(n = 15,
size = 5,
prob = 0.1))
crd_bin <- gexp(mu = 20,
err = e_binom,
r = 5,
fe = list(f1 = c(1, 4, 1)))
summary(crd_bin)
mod <- aov(Y1 ~ X1,
data = crd_bin$dfm)
shapiro.test(mod$res)
#! Factorial Experiment (FE) - CRD - QL
fe_crd00 <- gexp(mu = 0,
r = 2,
fe = list(f1 = c(1, 1, 5),
f2 = c(1, 1),
f3 = c(2, 2, 1)),
fl = list(A = paste('a',
1:3,
sep = ''),
B = paste('b',
1:2,
sep = ''),
C = paste('c',
1:3,
sep = '')),
round = 0,
type = 'FE')
fe_crd00$X
summary(fe_crd00)
#! Factorial Experiment (FE) - With interaction - CRD - QL
fe_crd01 <- gexp(mu = 30,
fe = list(f1 = c(1, 1, 3),
f2 = c(1, 1)),
fl = list(A = paste('a',
1:3,
sep = ''),
B = paste('b',
1:2,
sep = '')),
inte = c(3, 1, 1, 1, 1, 5), # (3*2)
round = 1,
type = 'FE')
summary(fe_crd01)
#! Split-plot Experiment (SPE) - CRD - QL
split_crd <- gexp(mu = 30,
fe = list(f1 = c(1, 1),
f2 = c(2, 3)),
fl = list(P = paste('p',
1:2,
sep = ''),
SP = paste('sp',
1:2,
sep = '')),
inte = c(1, 15, 1, 1), # (2*2)
round = 1,
type = 'SPE',
design = 'CRD')
split_crd$X
split_crd$Z
summary(split_crd)
split_crd01 <- gexp(mu = 30,
r = 3,
fe = list(f1 = c(1, 1),
f2 = c(2, 3),
f3 = c(1, 1, 1)),
fl = list(P = paste('p',
1:2,
sep = ''),
A = paste('a',
1:2,
sep = ''),
B = paste('b',
1:3,
sep = '')),
round = 1,
type = 'SPE',
design = 'CRD')
split_crd01$X
split_crd01$Z
summary(split_crd01)
#! Randomized Complete Block Design (RCBD) - QL
# 1 factor, 3 blocks
rcbd <- gexp(mu = 0,
r = 2,
fe = list(f1 = c(5, 1, 1)),
fl = list(TR = LETTERS[1:3]),
blke = c(1, 2, 3),
blkl = list(BLK = paste('B',
1:3,
sep = '')),
round = 1,
design = 'RCBD')
rcbd$X
summary(rcbd)
#! Factorial Experiment (FE) - RCBD - QL
fe_rcbd <- gexp(mu = 30,
r = 2,
fe = list(f1 = c(1, 1, 1),
f2 = c(2, 3)),
blke = c(1, 3),
inte = c(1, 15, 1, 1, 5, 1), # (3*2)
round = 1,
type = 'FE',
design = 'RCBD')
summary(fe_rcbd)
#! Multivariated - RCBD - QL
rcbd_m <- gexp(mu = c(0, 2),
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
1, 2,
1, 1),
ncol = 2,
byrow = TRUE),
round = 1,
design = 'RCBD')
summary(rcbd_m)
#! Split-plot Experiment (SPE) - RCBD - QL
split_rcbd <- gexp(mu = 30,
r = 2,
fe = list(f1 = c(1, 1),
f2 = c(2, 3),
f3 = c(1, 1, 1)),
fl = list(P = paste('p',
1:2,
sep = ''),
B = paste('b',
1:2,
sep = ''),
C = paste('c',
1:3,
sep = '')),
blke = c(1, 2),
blkl = list(BLK = paste('B',
1:2,
sep = '')),
round = 1,
type = 'SPE',
design = 'RCBD')
split_rcbd$Z
summary(split_rcbd)
#! Latin Square Design (LSD) - QL
#!. Warning!!!! r = 5 by default
lsd00 <- gexp(design = 'LSD')
#Set r = 1 to omiting warning
lsd01 <- gexp(mu = 30,
r = 1,
fe = list(f1 = c(1, 1, 10)),
rowe = c(1, 1, 1),
cole = c(1, 1, 1),
rowl = list(Row = paste('r',
1:3,
sep = '')),
coll = list(Col = paste('c',
1:3,
sep = '')),
round = 0,
design = 'LSD')
summary(lsd01)
#! Factorial Experiment (FE) - LSD - QL
fe_lsd <- gexp(mu = 30,
r = 1,
fe = list(f1 = c(1, 1),
f2 = c(2, 3)),
rowe = c(1, 3, 2, 1),
cole = c(2, 2, 1, 1),
rowl = list(Row = paste('r',
1:4,
sep = '')),
coll = list(Col = paste('c',
1:4,
sep = '')),
inte = c(1, 15, 1, 1), # (2*2)
round = 1,
type = 'FE',
design = 'LSD')
summary(fe_lsd)
#! Split-plot Experiment (SPE) - LSD - QL
split_lsd <- gexp(mu = 30,
r = 1,
fe = list(f1 = c(1, 1, 2),
f2 = c(2, 3, 1)),
fl = list(P = paste('p',
1:3,
sep = ''),
SP = paste('sp',
1:3,
sep = '')),
inte = c(1, 15, 1, 1, 1, 1, 1, 1, 1), # (3*3)
rowe = c(1, 1, 1),
cole = c(1, 1, 1),
rowl = list(Row = paste('r',
1:3,
sep = '')),
coll = list(Col = paste('c',
1:3,
sep = '')),
round = 1,
type = 'SPE',
design = 'LSD')
summary(split_lsd)
#!_____________________________
#! Quantitative Factor(s) (QT)
#!_____________________________
#! CRD - Orthogonal polynomials
# Linear effect
# Nonsense(error is 0)
# Default contrasts: Orthogonal contrasts
r <- 4 # (repet. number)
fln <- 5 # (factor levels number)
level <- c(0, 10, 20, 30, 40)
crd_lo <- gexp(mu = 1, #in this case, mu=beta0 (intercept)
r = r,
fe = list(f1 = c(2, 0, 0, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
err = matrix(0,
nrow = r*fln),
round = 2)
crd_lo$X
summary(crd_lo)
plot(Y1 ~ Dose,
crd_lo$dfm)
# Quadratic effect
crd_qo <- gexp(mu = 2,
r = r,
fe = list(f1 = c(0, 3, 0, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
err = matrix(0,
nrow = r*fln))
summary(crd_qo)
plot(Y1 ~ Dose,
crd_qo$dfm)
# Cubic effect
crd_co <- gexp(mu = 2,
r = r,
fe = list(f1 = c(1, 1, 3, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
err = matrix(0,
nrow = r*fln))
summary(crd_co)
plot(Y1 ~ Dose,
crd_co$dfm)
# Not orthogonal polynomials
# Linear
cont_crd <- matrix(c(level,
level^2,
level^3,
level^4),
ncol = 4)
crd_l <- gexp(mu = 2,
r = 2,
fe = list(f1 = c(10, 0, 0, 0)), #b1 #b2 #b3 #b4
fl = list(Dose = level),
contrasts = list(Dose = cont_crd))
crd_l$X
summary(crd_l)
plot(Y1 ~ Dose,
crd_l$dfm)
reg <- lm(Y1 ~ Dose + I(Dose^2) + I(Dose^3) + I(Dose^4),
data = crd_l$dfm)
summary(reg)
# Linear and quadratic
level1 <- seq(0,30,by = 10)
cont_crd1 <- matrix(c(level1,
level1^2,
level1^3),
ncol = 3)
level2 <- 1:4
cont_crd2 <- matrix(c(level2,
level2^2,
level2^3),
ncol = 3)
crd_lq <- gexp(mu = 1,
r = 2,
fe = list(f1 = c(10, 0, 0), #b1 #b2 #b3
f2 = c(1, 8, 0)),
fl = list(P = level1,
N = level2),
contrasts = list(N = cont_crd2,
P = cont_crd1))
crd_lq$X
summary(crd_lq)
with(crd_lq$dfm,
plot(Y1 ~ P))
with(crd_lq$dfm,
plot(Y1 ~ N))
# Multivariated
crd_m <- gexp(mu = c(2, 10),
r = 4,
fe = list(f1 = matrix(c(10, 0, #L Q
0, 10,
0, 0),
ncol = 2,
byrow = TRUE)),
fl = list(Dose = level1),
contrasts = list(Dose = cont_crd1))
with(crd_m$dfm,
plot(Y1 ~ Dose))
with(crd_m$dfm,
plot(Y2 ~ Dose))
# RCBD - Orthogonal polynomios
level3 <- c(0, 2, 4, 6)
rcbd <- gexp(mu = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
blke = c(1, 2, 3),
r = 2,
fl = list(Dose = level3),
blkl = list(Blk = c('B1', 'B2', 'B3')),
design = 'RCBD')
rcbd$X
summary(rcbd)
# Not orthogonal
cont_crd3 <- matrix(c(level3, level3^2, level3^3),
ncol = 3)
rcbd_01 <- gexp(mu = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
blke = c(1, 2, 3),
r = 2,
fl = list(Dose = level3),
blkl = list(Blk = c('B1', 'B2', 'B3')),
contrasts = list(Dose = cont_crd3),
design = 'RCBD')
rcbd_01$X
summary(rcbd_01)
# Orthogonal polynomios - LSD
lsd <- gexp(mu = 1,
r = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
rowe = rep(1, 4),
cole = rep(1, 4),
fl = list(Dose = level1),
design = 'LSD')
lsd$X
summary(lsd)
lsd_01 <- gexp(mu = 1,
r = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
rowe = rep(1, 4),
cole = rep(1, 4),
rowl = list(row = paste('r',
1:4,
sep = '')),
fl = list(Dose = level1),
design = 'LSD')
lsd_01$X
summary(lsd_01)
# Not orthogonal
lsd_02 <- gexp(mu = 1,
r = 1,
fe = list(f1 = c(3, 0, 0)), #b1 #b2 #b3
rowe = rep(1, 4),
cole = rep(1, 4),
fl = list(Dose = level3),
contrasts = list(Dose = cont_crd3),
design = 'LSD')
lsd_02$X
str(lsd_02)
#!__________________________________________________________________________
#! Hibrid: qualitative and quantitative factors in the same experiment - HB
#!__________________________________________________________________________
#! CRD - HB
r <- 2 # (repet. number)
fl1 <- 4# (first factor levels number)
fl2 <- 3# (second factor levels number)
crd_hb <- gexp(mu = 1, #in this case, mu=beta0 (intercept)
r = r,
fe = list(f1 = c(2, 0, 0), #b1 #b2 #b3
f2 = c(1, 1, 3)),
fl = list(Dose = seq(0,30,
by = 10),
Trat = LETTERS[1:3]),
err = matrix(0,
nrow = r*fl1*fl2),
round = 2)
crd_hb$X
summary(crd_hb)
#Only one contrasts!
crd_hb2 <- gexp(mu = 1, #in this case, mu=beta0 (intercept)
r = r,
fe = list(f1 = c(2, 0, 0), #b1 #b2 #b3
f2 = c(1, 1, 3)),
fl = list(Dose = level1,
Trat = LETTERS[1:3]),
err = matrix(0,
nrow = r*fl1*fl2),
contrasts = list(Dose = cont_crd1),
round = 2)
crd_hb2$X
summary(crd_hb)
#! RCBD - HB
r <- 2
blke <- c(1, 2)
level <- c(0, 10, 20, 30)
(error <- matrix(rep(0,
4^1*3^1*r*length(blke)),
ncol=1))
rcbd_hb <- gexp(mu = 2,
err = error,
r = r,
fe = list(f1 = c(0, 1, 0), # Qualitative
f2 = c(1, 0, 0)), # Quantitative linear
fl = list(Var = LETTERS[1:3],
Dose = level),
blke = blke,
blkl = list(Blk = c('B1', 'B2')),
design = 'RCBD')
rcbd_hb$X
summary(rcbd_hb)
str(rcbd_hb)
#! LSD - QT
set.seed(3)
lsd <- gexp(mu = 100,
r = 1,
fe = list(f1 = c(10, # b1
20, # b2
0, # b3
0)), # b4
fl = list(tra = seq(0,
40,
by = 10)),
rowe = c(1, 2, 3, 4, 5),
rowl = list(row = paste('r',
1:5,
sep = '')),
cole = c(5, 4, 3, 2, 1),
coll = list(col = paste('c',
1:5,
sep = '')),
design = 'LSD')
summary(lsd)
plot(Y1 ~ tra, lsd$dfm)
#! FE - LSD - QT
fe_lsd <- gexp(mu = 10,
fe = list(f1 = c(2, 3),
f2 = c(5, # b1*
0, # b2
0, # b3
0)), # b4
rowe = rep(1, 10),
cole = rep(1, 10),
fl = list(var = paste('v',
1:2,
sep = ''),
tra = seq(0,
40,
by = 10)),
coll = list(col = paste('c',
1:10,
sep = '')),
rowl = list(row = paste('r',
1:10,
sep = '')),
type = 'FE',
design = 'LSD')
fe_lsd$X
str(fe_lsd)
summary(fe_lsd)
plot(Y1 ~ tra,
fe_lsd$dfm)
#! SPE - QL - QT
spe_lsd <- gexp(mu = 100,
r = 1,
fe = list(f1 = c(2, 3, 1),
f2 = c(1, # b1
5, # b2*
1)), # b3
fl = list(p = paste('p',
1:3,
sep = ''),
sp = seq(0,
30,
by = 10)),
rowe = c(1, 2, 3),
cole = c(3, 2, 1),
rowl = list(row = paste('r',
1:3,
sep = '')),
coll = list(col = paste('c',
1:3,
sep = '')),
round = 1,
type = 'SPE',
design = 'LSD')
summary(spe_lsd)
plot(spe_lsd)
#++++++++++++++++++++++++++++#
# MULTIVARIATE APPROACH #
#++++++++++++++++++++++++++++#
#! CRD - QL
# Error = 0 - Nonsense (you can easily undertand the effects)
r <- 2 # (repet. number)
fln <- 3 # (factor levels number)
crd_m01 <- gexp(mu = c(0,10),
r = r,
fe = list(f1 = matrix(c(1, 0, #Y1 Y2
2, 1,
3, 3),
ncol = 2,
byrow = TRUE)),
err = mvtnorm::rmvnorm(n = fln * r,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
round = 0)
summary(crd_m01)
#! FE - CRD - QL
r <- 2
crd_m_fe01 <- gexp(mu = c(0, 0),
r = r,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * r,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
type = 'FE',
round = 1)
summary(crd_m_fe01)
#! FE - CRD - QL
# Using default error
set.seed(30)
crd_m_fe02 <- gexp(mu = c(0, 2),
r = 3,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
type = 'FE',
round = 1)
summary(crd_m_fe02)
#! SPE - CRD - QL
# Using default error
crd_m_spe01 <- gexp(mu = c(0, 2),
r = 3,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
type = 'SPE',
round = 1)
summary(crd_m_spe01)
#! RCBD - QL
r <- 2 # (repet. number)
fln <- 3 # (factor levels number)
bln <- 3 # (block levels number)
rcbd_m01 <- gexp(mu = c(0,10),
r = r,
fe = list(f1 = matrix(c(1, 0, #Y1 Y2
2, 1,
3, 3),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
err = mvtnorm::rmvnorm(n = fln * r * bln,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
design = 'RCBD',
round = 0)
summary(rcbd_m01)
#! FE - RCBD - QL
rcbd_m_fe01 <- gexp(mu = c(0, 0),
r = r,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * r * bln,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'FE',
design = 'RCBD',
round = 1)
summary(rcbd_m_fe01)
#! SPE - RCBD - QL
rcbd_m_spe01 <- gexp(mu = c(0, 2),
r = 2,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE),
f3 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'SPE',
design = 'RCBD',
round = 1)
summary(rcbd_m_spe01)
#! LSD - QL
lsd_m01 <- gexp(mu = c(0,10),
r = 1,
fe = list(f1 = matrix(c(1, 0,
2, 1,
3, 3),
ncol = 2,
byrow = TRUE)),
rowe = matrix(rep(1, 6),
ncol = 2),
cole = matrix(rep(1, 6),
ncol = 2),
err = mvtnorm::rmvnorm(n = 3^2,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
design = 'LSD',
round = 0)
summary(lsd_m01)
#! LSD/FE - QL
lsd_m_fe01 <- gexp(mu = c(0, 0),
r = 1,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * 6,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
#Y1 Y2
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
#Y1 Y2
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
rowe = matrix(rep(1, 12),
ncol = 2),
cole = matrix(rep(1, 12),
ncol = 2),
type = 'FE',
design = 'LSD',
round = 1)
summary(lsd_m_fe01)
#! SPE - LSD - QL
# Using default error
lsd_m_spe01 <- gexp(mu = c(0, 2),
r = 1,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
rowe = matrix(rep(1, 6),
ncol = 2),
cole = matrix(rep(1, 6),
ncol = 2),
type = 'SPE',
design = 'LSD',
round = 1)
summary(lsd_m_spe01)
#! FE - RCBD - QL
r = 1
bln = 3
fe_rcbd_m <- gexp(mu = c(0, 0),
r = 1,
err = mvtnorm::rmvnorm(n = 3^1 * 2^1 * r * bln,
sigma = matrix(c(0, 0,
0, 0),
ncol = 2)),
fe = list(f1 = matrix(c(0, 3, #X1 X1
1, 4, #X2 X2
2, 5), #X3 X3
ncol = 2,
byrow = TRUE),
f2 = matrix(c(0, 2, #X1 X1
1, 3), #X2 X2
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'FE',
design = 'RCBD')
str(fe_rcbd_m)
summary(fe_rcbd_m)
#! SPE - RCBD - QL
spe_rcbd_m <- gexp(mu = c(0, 2),
r = 3,
fe = list(f1 = matrix(c(1, 1,
5, 1,
1, 1),
ncol = 2,
byrow = TRUE),
f2 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE),
f3 = matrix(c(1, 3,
2, 2),
ncol = 2,
byrow = TRUE)),
blke = matrix(c(2, 1,
4, 1,
6, 1),
ncol = 2,
byrow = TRUE),
type = 'SPE',
design = 'RCBD')
str(spe_rcbd_m)
summary(spe_rcbd_m)
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.