CI_contrast: Constructing confidence regions for specific treatment effect...

In rwang14/implant: A High-throughput Phenotyping Pipeline for Image Processing and Functional Growth Curve Analysis

Description

This is a special case of the function "CI" in this package. This function is easier to use when studying the difference between two specific treatments/genotype, without specifying the linear combination needed in "CI".

Usage

CI_contrast(fit, j1, j2, alpha)

Arguments

fit	An object of output obtained by function "fanova".
j1	A positive integer, specifying the columns of the design matrix corresponding to one treatment of interest. Users can use the design matrix output from the function fanova( ), to idenfity the value of j1. See the example part for details.
j2	A positive integer different from j1. It specifies the columns of the design matrix corresponding to the other treatment of interest. Users can use the design matrix output from the function fanova( ), to idenfity the value of j2. See the example part for details.
alpha	A positive small number between 0 and 1. 1-alpha gives the confidence level. In default, alpha = 0.05.

Details

We can test the significance of the treatment effects of interest by constructing the corresponding confidence regions with the function ``CI\_contrast (fit, j1, j2, …)", where fit is the output from "fanova ( )", and j1 and j2 specify the columns of the design matrix corresponding to the treatments of interest. This is the simplified version of “'CI\_contrast" by setting the j1th and j2th elements of L being 1 and -1, and all other elements beign 0.

Value

trt	a t by 1 vector, which refers to the estimated treatment effect, where t is the number of observation time points specified in the argument tt in the function fanova( ).
lb	a t by 1 vector, which refers to the lower bound of the (1-alpha) confidence band, where t is the number of observation time points specified in the argument tt in the function fanova( ).
ub	a t by 1 vector, which refers to the the upper bound of the (1-alpha) confidence band, where t is the number of observation time points specified in the argument tt in the function fanova( ).

See Also

fanova

Examples

#load the data
data_new = read.csv(system.file("extdata","data.csv",
                                 package = "implant", mustWork = TRUE))
#The first three columns of data_new refer to the original position of the observations from the
#original dataset, genotype and block, respectively
Y = data_new[,-c(1:3)]
#This step is to factorize each factor
Genotype = as.factor(data_new$Genotype)
Block = as.factor(data_new$Block)
X = data.frame(Genotype,Block)
formula = "~ Genotype + Block"
tt = seq(from = 0, to = 44,by = 2)
#fit anova
fit = fanova(Y.na.mat = Y, X = X, tt = tt, formula, K.int = 6, order = 4, lower = -10, upper = 15)
fit$design_mat
> fit$design_mat
       (Intercept) Genotype2  Genotype  Block2  Block3
1           1          0           1      0      0
2           1          0           0      0      0
3           1          1           0      0      0
4           1          1           0      1      0
5           1          0           1      1      0
6           1          0           0      1      0
7           1          0           1      0      1
8           1          1           0      0      1
9           1          0           0      0      1
#We want to test the significance between block 2 and block 1,
#that is, we are interested in the 1st, the 4th column of the design matrix.
#Therefore, we can define:
ci_diff = CI_contrast(fit = fit, j1 = 4, j2 = 1, alpha = 0.05)
plot(tt,ci_diff$trt,type = "l")
lines(tt,ci_diff$lb, col = "blue")
lines(tt,ci_diff$ub, col = "blue")

rwang14/implant documentation built on Sept. 6, 2020, 3:21 a.m.