| gexp {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
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)