Segmentation of Partial Least Square Structural Equation Modeling Using Kernel K-Means Clustering (PLS SEM KKC) is the development of a new method in PLS SEM segmentation that uses non-linear clustering methods. Segmentation is carried out based on residual values of measurement and structural PLS SEM modeling which often exhibit non-linear separability and thus require a non-linear separable, thus requiring a non-linear PLS SEM segmentation. Main contribution of this study is the integration of kernel-based clustering into PLS SEM segmentation. Method effectively addresses unobserved heterogeneity by capturing non-linear residual patterns, leading to more accurate modelsSegmentation of Partial Least Square Structural Equation Modeling Using Kernel K-Means Clustering (PLS SEM KKC) is the development of a new method in PLS SEM segmentation that uses non-linear clustering methods. Segmentation is carried out based on residual values of measurement and structural PLS SEM modeling which often exhibit non-linear separability and thus require a non-linear separable, thus requiring a non-linear PLS SEM segmentation. Main contribution of this study is the integration of kernel-based clustering into PLS SEM segmentation. Method effectively addresses unobserved heterogeneity by capturing non-linear residual patterns, leading to more accurate models.
This package is an extension of the ResiPLS package, which develops SEMPLS analysis with segmentation using Kernel K-Means. If you are interested in this package, please contact me at (cindy.cahyaning@umsida.ac.id).
The following are the functions built into the SEMPLSKKC Package:
import_dari_excel = function(file_path="data.xlsx") {
library(readxl)
Variabel=read_excel(file_path, sheet = "Variabel")
Variabel=data.frame(Variabel=Variabel$Variabel)
Indikator=read_excel(file_path, sheet = "Indikator")
Indikator=data.frame(Indikator=Indikator$Indikator)
Data=read_excel(file_path, sheet = "Data")
MIM=read_excel(file_path, sheet = "MIM")
MOM=read_excel(file_path, sheet = "MOM")
return(list(Variabel = Variabel, Indikator = Indikator, Data = Data, MIM = MIM, MOM = MOM))
}
ols = function(X, y) {
beta = solve(t(X) %*% X) %*% t(X) %*% y
return(beta)
}
plsalgorithm <- function(data, matrix_outer_model, matrix_inner_model, inner_weight_scheme="Centroid", indicator, laten_variables){
#Set Output
outputplsalgoritm=list()
indexoutput=0
#Counting the number of indicators
n_indicator=nrow(matrix_outer_model)
#Counting the number of laten variable
n_laten_variables=ncol(matrix_outer_model)
#Counting the number of observations
n=nrow(data)
#Data standardization in matrix form
standardize_data=scale(data)
#Mapping indicator to latent variable
indicator_to_laten_variable_map=list()
for (i in 1:n_laten_variables) {
indicator_to_laten_variable_connection=c()
for (j in 1:n_indicator) {
if (matrix_outer_model[j,i]==1 | matrix_outer_model[j,i]==2 | matrix_outer_model[j,i]==3) {
indicator_to_laten_variable_connection=append(indicator_to_laten_variable_connection,j)
}
}
indicator_to_laten_variable_map=append(indicator_to_laten_variable_map,list(indicator_to_laten_variable_connection))
}
#Outer model type mapping
type_outer_model_map=list()
for (i in 1:n_laten_variables) {
type_outer_model=c()
for (j in 1:n_indicator) {
if (matrix_outer_model[j,i]==1 | matrix_outer_model[j,i]==2 | matrix_outer_model[j,i]==3) {
type_outer_model=append(type_outer_model,matrix_outer_model[j,i])
}
}
type_outer_model_map=append(type_outer_model_map,list(type_outer_model))
}
#Standardization of data in list form
list_indicator=list()
for (i in 1:n_laten_variables) {
part_standardize_data=standardize_data[,indicator_to_laten_variable_map[[i]]]
list_indicator=append(list_indicator,list(part_standardize_data))
}
#Initialization Weight
Weight=list()
for (i in 1:n_laten_variables) {
vector_one=rep(1,length(indicator_to_laten_variable_map[[i]]))
matrix_one=matrix(vector_one,nrow = length(indicator_to_laten_variable_map[[i]]),ncol = 1)
Weight=append(Weight,list(matrix_one))
}
#Counting f
f=list()
for (i in 1:n_laten_variables) {
f_i=sqrt(1/(t(Weight[[i]])%*%var(list_indicator[[i]])%*%Weight[[i]]))
f=append(f,list(f_i))
}
#Counting Outer Weight
outer_weight=list()
for (i in 1:n_laten_variables) {
outer_weight_i=Weight[[i]]%*%f[[i]]
outer_weight=append(outer_weight,list(outer_weight_i))
}
#Outsite Approximation
Y=list()
for (i in 1:n_laten_variables) {
Y_i=list_indicator[[i]]%*%outer_weight[[i]]
Y=append(Y,list(Y_i))
}
#Outer Model Iteration
check_result = FALSE
iteration=1
while (check_result <= FALSE) {
## step 1 inner weight
v=matrix(0,nrow = n_laten_variables,ncol = n_laten_variables)
if (inner_weight_scheme == "Centroid") {
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (matrix_inner_model[i,j] == 1) {
v[i,j]=cor(Y[[i]],Y[[j]])/abs(cor(Y[[i]],Y[[j]]))
v[j,i]=cor(Y[[i]],Y[[j]])/abs(cor(Y[[i]],Y[[j]]))
}
}
}
} else if (inner_weight_scheme == "Factor") {
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (matrix_inner_model[i,j] == 1) {
v[i,j]=cor(Y[[i]],Y[[j]])
v[j,i]=cor(Y[[i]],Y[[j]])
}
}
}
} else if (inner_weight_scheme == "Path") {
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (matrix_inner_model[i,j] == 1) {
v[i,j]=ols(Y[[i]],Y[[j]])
v[j,i]=cor(Y[[i]],Y[[j]])
}
}
}
}
#step 2 inside approximation
Y_ins=list()
for (i in 1:n_laten_variables) {
Y_ins_i=matrix(0,nrow = n,ncol = 1)
for (j in 1:n_laten_variables) {
Y_ins_i=Y_ins_i+Y[[j]]%*%v[i,j]
}
Y_ins=append(Y_ins,list(Y_ins_i))
}
#step 3 outer weight
##weight
weight_new=list()
for (i in 1:n_laten_variables) {
# 1=single construct, 2=reflective, 3=formative
if (type_outer_model_map[[i]][1]==2) {
weight_new_i=matrix(0,nrow = length(type_outer_model_map[[i]]), ncol = 1)
for (j in 1:length(type_outer_model_map[[i]])) {
weight_new_i[j,1]=ols(Y_ins[[i]],list_indicator[[i]][,j])
}
weight_new=append(weight_new,list(weight_new_i))
} else if (type_outer_model_map[[i]][1]==3) {
weight_new_i=ols(list_indicator[[i]],Y_ins[[i]])
weight_new=append(weight_new,list(weight_new_i))
} else if (type_outer_model_map[[i]][1]==1) {
weight_new_i=Weight[[i]]
weight_new=append(weight_new,list(weight_new_i))
}
}
check_result = TRUE
for (i in 1:length(Weight)) {
difference=abs(Weight[[i]]-weight_new[[i]])
if(!all(difference<0.00001)){
check_result = FALSE
break
}
}
##Adding iteration value
iteration=iteration+1
for (i in 1:n_laten_variables) {
Weight[[i]]=weight_new[[i]]
}
##Counting f
f=list()
for (i in 1:n_laten_variables) {
f_i=sqrt(1/(t(Weight[[i]])%*%var(list_indicator[[i]])%*%Weight[[i]]))
f=append(f,list(f_i))
}
##Counting outer weight
outer_weight=list()
for (i in 1:n_laten_variables) {
outer_weight_i=Weight[[i]]%*%f[[i]]
outer_weight=append(outer_weight,list(outer_weight_i))
}
#step 4 (outside approximation)
Y=list()
for (i in 1:n_laten_variables) {
Y_i=list_indicator[[i]]%*%outer_weight[[i]]
Y=append(Y,list(Y_i))
}
}
# Outer Weight Final and Skor Laten Final
cat("Number of Iterasi:",iteration,"\n")
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=outer_weight
new_text=paste("Outer Weight Final")
names(outputplsalgoritm)[indexoutput]=new_text
cat("Outer Weight","\n")
for(i in seq_along(outer_weight)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(outer_weight[[i]])
cat("\n")
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=Y
new_text=paste("Y Final")
names(outputplsalgoritm)[indexoutput]=new_text
cat("Skor Variabel Laten","\n")
for(i in seq_along(Y)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(head(Y[[i]]))
cat("\n")
}
#stage 2
##Creating a list of exogenous columns
list_exogenous_columns=list()
for (i in 1:n_laten_variables) {
exogenous_columns=c()
for (j in 1:n_laten_variables) {
if (matrix_inner_model[j,i]==1) {
exogenous_columns=append(exogenous_columns,j)
}
}
list_exogenous_columns=append(list_exogenous_columns,list(exogenous_columns))
}
##Eksogen
exogenous=list()
for (i in 1:n_laten_variables) {
k=length(list_exogenous_columns[[i]])
if (k==0) {
exogenous_i = NULL
} else if (k==1) {
fill=list_exogenous_columns[[i]][1]
exogenous_i = Y[[fill]]
} else if (k>1) {
for (j in 1:k) {
fill=list_exogenous_columns[[i]][j]
# print(fill)
if (j==1) {
exogenous_i = Y[[fill]]
} else {
exogenous_i = cbind(exogenous_i,Y[[fill]])
}
}
}
exogenous=append(exogenous,list(exogenous_i))
}
path_coefficient=list()
for (i in 1:n_laten_variables) {
if (is.null(list_exogenous_columns[[i]])) {
path_coefficient_i=NULL
} else {
path_coefficient_i=ols(exogenous[[i]],Y[[i]])
}
path_coefficient=append(path_coefficient,list(path_coefficient_i))
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=path_coefficient
names(outputplsalgoritm)[indexoutput]="Path Coefficient"
cat("Koefisien Jalur:\n ")
for(i in seq_along(path_coefficient)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(path_coefficient[[i]])
cat("\n")
}
##Outer Loading
outer_loading=list()
for (i in 1:n_laten_variables) {
if (type_outer_model_map[[i]][1]==2 | type_outer_model_map[[i]][1]==3) {
outer_loading_i=matrix(0,nrow = length(type_outer_model_map[[i]]), ncol = 1)
for (j in 1:length(type_outer_model_map[[i]])) {
outer_loading_i[j,1]=ols(Y[[i]],list_indicator[[i]][,j])
}
outer_loading=append(outer_loading,list(outer_loading_i))
} else if (type_outer_model_map[[i]][1]==1) {
outer_loading_i=1
outer_loading=append(outer_loading,list(outer_loading_i))
}
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=outer_loading
names(outputplsalgoritm)[indexoutput]="Outer Loading"
cat("Outer Loading:\n ")
for(i in seq_along(outer_loading)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(outer_loading[[i]])
cat("\n")
}
vector_outer_loading=unlist(outer_loading)
##Inner Residual
pertama=TRUE
for (i in 1:n_laten_variables) {
if (is.null(list_exogenous_columns[[i]])) {
} else {
innerresidual_i=Y[[i]]-(exogenous[[i]]%*%path_coefficient[[i]])
if (pertama==TRUE) {
innerresidual=matrix(innerresidual_i)
pertama=FALSE
} else if (pertama==FALSE) {
innerresidual=cbind(innerresidual,innerresidual_i)
}
}
}
##Outer Residual
pertama=TRUE
for (i in 1:n_laten_variables) {
for (j in 1:n_indicator) {
if (matrix_outer_model[j,i]==2 | matrix_outer_model[j,i]==3) {
outerresidual_i=standardize_data[,j]-(Y[[i]]%*%vector_outer_loading[j])
if (pertama==TRUE) {
outerresidual=matrix(outerresidual_i)
pertama=FALSE
} else if (pertama==FALSE) {
outerresidual=cbind(outerresidual,outerresidual_i)
}
}
}
}
##Matriks Residual (Combination of Outer Residual and Inner Residual)
matrixresidual=cbind(outerresidual,innerresidual)
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=matrixresidual
names(outputplsalgoritm)[indexoutput]="Matrix Residual"
cat("Matriks Residual Gabungan:\n ")
print(head(matrixresidual))
cat("\n")
##Composite Reliability (CR)
CR=list()
for (i in 1:length(outer_loading)) {
if (type_outer_model_map[[i]][1]==2) {
squared_loadings=outer_loading[[i]]^2
squared_sum_loadings=sum(outer_loading[[i]])^2
var_error=1-squared_loadings
sum_var_error=sum(var_error)
CR_i=squared_sum_loadings/(squared_sum_loadings+sum_var_error)
CR=append(CR,CR_i)
} else if (type_outer_model_map[[i]][1]==1 | type_outer_model_map[[i]][1]==3) {
CR_i=NA
CR=append(CR,CR_i)
}
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=CR
names(outputplsalgoritm)[indexoutput]="CR"
cat("CR:\n ")
for(i in seq_along(CR)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(CR[[i]])
cat("\n")
}
##Average Variance Extracted(AVE)
AVE=list()
for (i in 1:length(outer_loading)) {
if (type_outer_model_map[[i]][1]==2) {
squared_loadings=outer_loading[[i]]^2
sum_squared_loadings=sum(squared_loadings)
var_error=1-squared_loadings
sum_var_error=sum(var_error)
AVE_i=sum_squared_loadings/(sum_squared_loadings+sum_var_error)
AVE=append(AVE,AVE_i)
} else if (type_outer_model_map[[i]][1]==1 | type_outer_model_map[[i]][1]==3) {
AVE_i=NA
AVE=append(AVE,AVE_i)
}
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=AVE
names(outputplsalgoritm)[indexoutput]="AVE"
cat("AVE:\n ")
for(i in seq_along(AVE)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(AVE[[i]])
cat("\n")
}
##Heterotrait-Monotrait Ratio (HTMT)
htmt_matrix = matrix(NA, n_laten_variables, n_laten_variables)
rownames(htmt_matrix)=laten_variables$Variabel
colnames(htmt_matrix)=laten_variables$Variabel
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (i != j) {
indicators_i=list_indicator[[i]]
indicators_j=list_indicator[[j]]
cross_correlations=cor(indicators_i, indicators_j)
within_correlations_i=cor(indicators_i)
within_correlations_j=cor(indicators_j)
mean_cross_correlation=mean(abs(cross_correlations))
mean_within_correlation_i=mean(abs(within_correlations_i[lower.tri(within_correlations_i)]))
mean_within_correlation_j=mean(abs(within_correlations_j[lower.tri(within_correlations_j)]))
htmt_value=mean_cross_correlation/sqrt(mean_within_correlation_i*mean_within_correlation_j)
htmt_matrix[i, j]=htmt_value
}
}
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=htmt_matrix
names(outputplsalgoritm)[indexoutput]="HTMT Matrix"
cat("HTMT Matrix:\n ")
print(htmt_matrix)
cat("\n ")
##R Square
R_Square=list()
for (i in 1:n_laten_variables) {
if (!is.null(exogenous[[i]]) && !is.null(path_coefficient[[i]])) {
y_actual = Y[[i]]
y_predicted = exogenous[[i]] %*% path_coefficient[[i]]
ss_total = sum((y_actual - mean(y_actual))^2)
ss_residual = sum((y_actual - y_predicted)^2)
r_square = 1 - (ss_residual / ss_total)
R_Square[[i]] = r_square
} else {
R_Square[[i]] = NA
}
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=R_Square
names(outputplsalgoritm)[indexoutput]="R Square"
cat("R Square:\n ")
for(i in seq_along(R_Square)) {
cat("Variabel Laten ",laten_variables[i,1],"\n")
print(R_Square[[i]])
cat("\n")
}
#output
return(outputplsalgoritm)
}
plsalgorithm_bootstrap <- function(data, matrix_outer_model, matrix_inner_model, inner_weight_scheme="Centroid", indicator, laten_variables, n_bootstrap=1000) {
# Function to calculate PLS Algorithm without bootstrapping
pls_algorithm_core <- function(data) {
plsalgoritmlite <- function(data, matrix_outer_model, matrix_inner_model, inner_weight_scheme="Centroid", indicator, laten_variables){
#Set Output
outputplsalgoritm=list()
indexoutput=0
#Counting the number of indicators
n_indicator=nrow(matrix_outer_model)
#Counting the number of laten variables
n_laten_variables=ncol(matrix_outer_model)
#Counting the number of observation
n=nrow(data)
#Data standardization in matrix form
standardize_data=scale(data)
#Mapping indicator to latent variable
indicator_to_laten_variable_map=list()
for (i in 1:n_laten_variables) {
indicator_to_laten_variable_connection=c()
for (j in 1:n_indicator) {
if (matrix_outer_model[j,i]==1 | matrix_outer_model[j,i]==2 | matrix_outer_model[j,i]==3) {
indicator_to_laten_variable_connection=append(indicator_to_laten_variable_connection,j)
}
}
indicator_to_laten_variable_map=append(indicator_to_laten_variable_map,list(indicator_to_laten_variable_connection))
}
#Outer model type mapping
type_outer_model_map=list()
for (i in 1:n_laten_variables) {
type_outer_model=c()
for (j in 1:n_indicator) {
if (matrix_outer_model[j,i]==1 | matrix_outer_model[j,i]==2 | matrix_outer_model[j,i]==3) {
type_outer_model=append(type_outer_model,matrix_outer_model[j,i])
}
}
type_outer_model_map=append(type_outer_model_map,list(type_outer_model))
}
#Standardization of data in list form
list_indicator=list()
for (i in 1:n_laten_variables) {
part_standardize_data=standardize_data[,indicator_to_laten_variable_map[[i]]]
list_indicator=append(list_indicator,list(part_standardize_data))
}
# Initialization Weight
Weight=list()
for (i in 1:n_laten_variables) {
vector_one=rep(1,length(indicator_to_laten_variable_map[[i]]))
matrix_one=matrix(vector_one,nrow = length(indicator_to_laten_variable_map[[i]]),ncol = 1)
Weight=append(Weight,list(matrix_one))
}
#Counting nilai f
f=list()
for (i in 1:n_laten_variables) {
f_i=sqrt(1/(t(Weight[[i]])%*%var(list_indicator[[i]])%*%Weight[[i]]))
f=append(f,list(f_i))
}
#Counting Outer Weight
outer_weight=list()
for (i in 1:n_laten_variables) {
outer_weight_i=Weight[[i]]%*%f[[i]]
outer_weight=append(outer_weight,list(outer_weight_i))
}
#Outsite Approximation
Y=list()
for (i in 1:n_laten_variables) {
Y_i=list_indicator[[i]]%*%outer_weight[[i]]
Y=append(Y,list(Y_i))
}
#Step 1 Iteration Outer Model
check_result = FALSE
iteration=1
while (check_result <= FALSE) {
## step 1 inner weight
v=matrix(0,nrow = n_laten_variables,ncol = n_laten_variables)
if (inner_weight_scheme == "Centroid") {
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (matrix_inner_model[i,j] == 1) {
v[i,j]=cor(Y[[i]],Y[[j]])/abs(cor(Y[[i]],Y[[j]]))
v[j,i]=cor(Y[[i]],Y[[j]])/abs(cor(Y[[i]],Y[[j]]))
}
}
}
} else if (inner_weight_scheme == "Factor") {
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (matrix_inner_model[i,j] == 1) {
v[i,j]=cor(Y[[i]],Y[[j]])
v[j,i]=cor(Y[[i]],Y[[j]])
}
}
}
} else if (inner_weight_scheme == "Path") {
for (i in 1:n_laten_variables) {
for (j in 1:n_laten_variables) {
if (matrix_inner_model[i,j] == 1) {
v[i,j]=ols(Y[[i]],Y[[j]])
v[j,i]=cor(Y[[i]],Y[[j]])
}
}
}
}
#step 2 inside approximation
Y_ins=list()
for (i in 1:n_laten_variables) {
Y_ins_i=matrix(0,nrow = n,ncol = 1)
for (j in 1:n_laten_variables) {
Y_ins_i=Y_ins_i+Y[[j]]%*%v[i,j]
}
Y_ins=append(Y_ins,list(Y_ins_i))
}
#step 3 outer weight
##weight
weight_new=list()
for (i in 1:n_laten_variables) {
# 1=single construct, 2=reflective, 3=formative
if (type_outer_model_map[[i]][1]==2) {
weight_new_i=matrix(0,nrow = length(type_outer_model_map[[i]]), ncol = 1)
for (j in 1:length(type_outer_model_map[[i]])) {
weight_new_i[j,1]=ols(Y_ins[[i]],list_indicator[[i]][,j])
}
weight_new=append(weight_new,list(weight_new_i))
} else if (type_outer_model_map[[i]][1]==3) {
weight_new_i=ols(list_indicator[[i]],Y_ins[[i]])
weight_new=append(weight_new,list(weight_new_i))
} else if (type_outer_model_map[[i]][1]==1) {
weight_new_i=Weight[[i]]
weight_new=append(weight_new,list(weight_new_i))
}
}
check_result = TRUE
for (i in 1:length(Weight)) {
difference=abs(Weight[[i]]-weight_new[[i]])
if(!all(difference<0.00001)){
check_result = FALSE
break
}
}
iteration=iteration+1
for (i in 1:n_laten_variables) {
Weight[[i]]=weight_new[[i]]
}
##Counting f
f=list()
for (i in 1:n_laten_variables) {
f_i=sqrt(1/(t(Weight[[i]])%*%var(list_indicator[[i]])%*%Weight[[i]]))
f=append(f,list(f_i))
}
##Counting outer weight
outer_weight=list()
for (i in 1:n_laten_variables) {
outer_weight_i=Weight[[i]]%*%f[[i]]
outer_weight=append(outer_weight,list(outer_weight_i))
}
#step 4 (outside approximation)
Y=list()
for (i in 1:n_laten_variables) {
Y_i=list_indicator[[i]]%*%outer_weight[[i]]
Y=append(Y,list(Y_i))
}
}
#Print Final Outer Weight and Final Latent Score
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=outer_weight
new_text=paste("Outer Weight Final")
names(outputplsalgoritm)[indexoutput]=new_text
#step 2
##Creating a list of exogenous columns
list_exogenous_columns=list()
for (i in 1:n_laten_variables) {
exogenous_columns=c()
for (j in 1:n_laten_variables) {
if (matrix_inner_model[j,i]==1) {
exogenous_columns=append(exogenous_columns,j)
}
}
list_exogenous_columns=append(list_exogenous_columns,list(exogenous_columns))
}
##Eksogen
exogenous=list()
for (i in 1:n_laten_variables) {
k=length(list_exogenous_columns[[i]])
if (k==0) {
exogenous_i = NULL
} else if (k==1) {
fill=list_exogenous_columns[[i]][1]
exogenous_i = Y[[fill]]
} else if (k>1) {
for (j in 1:k) {
fill=list_exogenous_columns[[i]][j]
# print(fill)
if (j==1) {
exogenous_i = Y[[fill]]
} else {
exogenous_i = cbind(exogenous_i,Y[[fill]])
}
}
}
exogenous=append(exogenous,list(exogenous_i))
}
path_coefficient=list()
for (i in 1:n_laten_variables) {
if (is.null(list_exogenous_columns[[i]])) {
path_coefficient_i=NULL
} else {
path_coefficient_i=ols(exogenous[[i]],Y[[i]])
}
path_coefficient=append(path_coefficient,list(path_coefficient_i))
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=path_coefficient
names(outputplsalgoritm)[indexoutput]="Path Coefficient"
##Outer Loading
outer_loading=list()
for (i in 1:n_laten_variables) {
if (type_outer_model_map[[i]][1]==2 | type_outer_model_map[[i]][1]==3) {
outer_loading_i=matrix(0,nrow = length(type_outer_model_map[[i]]), ncol = 1)
for (j in 1:length(type_outer_model_map[[i]])) {
outer_loading_i[j,1]=ols(Y[[i]],list_indicator[[i]][,j])
}
outer_loading=append(outer_loading,list(outer_loading_i))
} else if (type_outer_model_map[[i]][1]==1) {
outer_loading_i=1
outer_loading=append(outer_loading,list(outer_loading_i))
}
}
indexoutput=indexoutput+1
outputplsalgoritm[[indexoutput]]=outer_loading
names(outputplsalgoritm)[indexoutput]="Outer Loading"
#output
return(outputplsalgoritm)
}
result <- plsalgoritmlite(data, matrix_outer_model, matrix_inner_model, inner_weight_scheme, indicator, laten_variables)
return(result)
}
# Calculate PLS algorithm for original data
result_original <- pls_algorithm_core(data)
# Initialization to save bootstrapping results
outer_weight_bootstrap <- list()
outer_loading_bootstrap <- list()
path_coefficient_bootstrap <- list()
# bootstrapping
for (i in 1:n_bootstrap) {
data_bootstrap <- data[sample(1:nrow(data), replace=TRUE), ]
result_bootstrap <- pls_algorithm_core(data_bootstrap)
if (i == 1) {
outer_weight_bootstrap <- result_bootstrap$`Outer Weight Final`
outer_loading_bootstrap <- result_bootstrap$`Outer Loading`
path_coefficient_bootstrap <- result_bootstrap$`Path Coefficient`
} else {
for (j in 1:nrow(laten_variables)) {
# print(outer_weight_bootstrap[[j]])
# print(result_bootstrap$`Outer Weight Final`[[j]])
if (!is.null(outer_weight_bootstrap[[j]]) ) {
outer_weight_bootstrap[[j]] <- cbind(outer_weight_bootstrap[[j]], result_bootstrap$`Outer Weight Final`[[j]])
}
if (!is.null(outer_loading_bootstrap[[j]]) ) {
outer_loading_bootstrap[[j]] <- cbind(outer_loading_bootstrap[[j]], result_bootstrap$`Outer Loading`[[j]])
}
if (!is.null(path_coefficient_bootstrap[[j]]) ) {
path_coefficient_bootstrap[[j]] <- cbind(path_coefficient_bootstrap[[j]], result_bootstrap$`Path Coefficient`[[j]])
}
}
}
}
#Calculating bootstrap statistics
howb=list()
for (j in 1:nrow(laten_variables)) {
if (!is.null(result_original$`Outer Weight Final`[[j]])) {
outer_weight_original=as.vector(result_original$`Outer Weight Final`[[j]])
average_outer_weight=rowMeans(outer_weight_bootstrap[[j]])
standar_error=as.vector(apply(outer_weight_bootstrap[[j]], 1, sd))
t_hitung=outer_weight_original/standar_error
p_value=2*(1-pnorm(abs(t_hitung)))
howb[[j]]=data.frame(outer_weight_original,
average_outer_weight,
standar_error,
t_hitung,
p_value)
} else {
outer_weight_original=NULL
average_outer_weight=NULL
standar_error=NULL
t_hitung=NULL
p_value=NULL
howb[[j]]=data.frame(outer_weight_original,
average_outer_weight,
standar_error,
t_hitung,
p_value)
}
}
holb=list()
for (j in 1:nrow(laten_variables)) {
if (!is.null(result_original$`Outer Loading`[[j]])) {
outer_loading_original=as.vector(result_original$`Outer Loading`[[j]])
average_outer_loading=rowMeans(outer_loading_bootstrap[[j]])
standar_error=as.vector(apply(outer_loading_bootstrap[[j]], 1, sd))
t_hitung=outer_loading_original/standar_error
p_value=2*(1-pnorm(abs(t_hitung)))
holb[[j]]=data.frame(outer_loading_original,
average_outer_loading,
standar_error,
t_hitung,
p_value)
} else {
outer_loading_original=NULL
average_outer_loading=NULL
standar_error=NULL
t_hitung=NULL
p_value=NULL
holb[[j]]=data.frame(outer_loading_original,
average_outer_loading,
standar_error,
t_hitung,
p_value)
}
}
hpcb=list()
for (j in 1:nrow(laten_variables)) {
if (!is.null(result_original$`Path Coefficient`[[j]])) {
path_coefficient_original=as.vector(result_original$`Path Coefficient`[[j]])
average_path_coefficient=rowMeans(path_coefficient_bootstrap[[j]])
standar_error=as.vector(apply(path_coefficient_bootstrap[[j]], 1, sd))
t_hitung=path_coefficient_original/standar_error
p_value=2*(1-pnorm(abs(t_hitung)))
hpcb[[j]]=data.frame(path_coefficient_original,
average_path_coefficient,
standar_error,
t_hitung,
p_value)
} else {
path_coefficient_original=NULL
average_path_coefficient=NULL
standar_error=NULL
t_hitung=NULL
p_value=NULL
hpcb[[j]]=data.frame(path_coefficient_original,
average_path_coefficient,
standar_error,
t_hitung,
p_value)
}
}
#Outer Weight
output <- list(
outer_weight=howb,
outer_loading=holb,
path_coefficient=hpcb
)
return(output)
}
euclidean_distance <- function(a, b) {
sqrt(sum((a - b)^2))
}
kmeans_custom <- function(data, k, max_iter = 100) {
n <- nrow(data)
d <- ncol(data)
# Randomly initialize centroids
set.seed(123)
centroids <- data[sample(1:n, k), ]
clusters <- rep(0, n)
for (iter in 1:max_iter) {
# Cluster assignment
for (i in 1:n) {
distances <- apply(centroids, 1, function(centroid) euclidean_distance(data[i, ], centroid))
clusters[i] <- which.min(distances)
}
# Save the previous centroid for convergence checking.
previous_centroids <- centroids
# Centroid update
for (j in 1:k) {
if (sum(clusters == j) > 0) {
centroids[j, ] <- colMeans(data[clusters == j, ])
}
}
# Convergence check
if (all(previous_centroids == centroids)) {
cat("Algoritma konvergen pada iterasi ke-", iter, "\n")
break
}
}
# Returns results in list form
return(list(clusters = clusters, centroids = centroids))
}
rbf_kernel <- function(x, y, sigma = 1) {
dist_sq <- sum((x - y)^2)
exp(-dist_sq / (2 * sigma^2))
}
create_kernel_matrix <- function(data, kernel_fun = rbf_kernel, sigma = 1) {
n <- nrow(data)
K <- matrix(0, n, n)
for (i in 1:n) {
for (j in 1:n) {
K[i, j] <- kernel_fun(data[i, ], data[j, ], sigma = sigma)
}
}
return(K)
}
kernel_kmeans <- function(data,
k,
max_iter = 100,
kernel_fun = rbf_kernel,
sigma = 1) {
# 1. Counting kernel matrix K (n x n)
K <- create_kernel_matrix(data, kernel_fun, sigma)
n <- nrow(data)
# 2. Random cluster initialization
set.seed(123)
clusters <- sample(1:k, n, replace = TRUE)
# Auxiliary variable to store the number of intracluster kernels
sum_K_in_cluster <- numeric(k)
sum_K_with_i <- matrix(0, nrow = k, ncol = n)
# -------------------------
# function updates the number of kernels for each cluster
# -------------------------
update_cluster_sums <- function(clusters) {
sum_K_in_cluster <<- numeric(k)
sum_K_with_i <<- matrix(0, nrow = k, ncol = n)
for (j in 1:k) {
idx_j <- which(clusters == j)
if (length(idx_j) > 0) {
# sum_{p,q in C_j} K[p, q]
sum_K_in_cluster[j] <<- sum(K[idx_j, idx_j])
# sum_{p in C_j} K[i, p]
for (i in 1:n) {
sum_K_with_i[j, i] <<- sum(K[i, idx_j])
}
}
}
}
# Inisialisasi sum kernel
update_cluster_sums(clusters)
# -------------------------
# Iteration Kernel K-Means
# -------------------------
for (iter in 1:max_iter) {
changed <- FALSE
for (i in 1:n) {
# Counting distance^2 in feature space
# distance^2 = K[i,i] - 2/|C_j| sum_{p in C_j} K[i,p] + 1/|C_j|^2 sum_{p,q in C_j} K[p,q]
dist_clusters <- numeric(k)
for (j in 1:k) {
idx_j <- which(clusters == j)
nj <- length(idx_j)
if (nj > 0) {
dist_clusters[j] <- K[i, i] -
(2 / nj) * sum_K_with_i[j, i] +
(1 / (nj^2)) * sum_K_in_cluster[j]
} else {
dist_clusters[j] <- Inf
}
}
new_cluster <- which.min(dist_clusters)
if (new_cluster != clusters[i]) {
clusters[i] <- new_cluster
changed <- TRUE
}
}
# Check convergence
if (!changed) {
cat("Kernel K-Means konvergen pada iterasi ke-", iter, "\n")
break
}
# Update sum-sum kernel
update_cluster_sums(clusters)
}
# Return results
return(list(
clusters = clusters
))
}To run this package, you can do this:
library(SEMPLSKKC)##
## Attaching package: 'SEMPLSKKC'## The following objects are masked _by_ '.GlobalEnv':
##
## create_kernel_matrix, euclidean_distance, import_dari_excel,
## kernel_kmeans, kmeans_custom, ols, plsalgorithm,
## plsalgorithm_bootstrap, rbf_kernelRunning SEMPLSKKC
inputdata=import_dari_excel(file_path = "DataSeminter.xlsx")
hasil=plsalgorithm(data = inputdata$Data,inner_weight_scheme = "Path",matrix_outer_model = inputdata$MOM, matrix_inner_model = inputdata$MIM, indicator = inputdata$Indikator, laten_variables = inputdata$Variabel)## Number of Iterasi: 7
## Outer Weight
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] 0.4078256
## [2,] 0.3948863
## [3,] 0.3516330
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] 0.4682900
## [2,] 0.3154480
## [3,] 0.3689983
##
## Variabel Laten Social Influence
## [,1]
## [1,] 0.7372356
## [2,] 0.4142433
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] 0.3906090
## [2,] 0.4973840
## [3,] 0.3232516
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.3231985
## [2,] 0.3223177
## [3,] 0.2944332
## [4,] 0.1980679
##
## Skor Variabel Laten
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] -0.193634
## [2,] -0.193634
## [3,] -0.193634
## [4,] -0.621821
## [5,] -0.193634
## [6,] -0.193634
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] -0.63653562
## [2,] 1.29973203
## [3,] 0.61976369
## [4,] -0.63653562
## [5,] 0.04343272
## [6,] -0.63653562
##
## Variabel Laten Social Influence
## [,1]
## [1,] -2.0094775
## [2,] -1.6434825
## [3,] -0.3188389
## [4,] 0.1604853
## [5,] -0.6848340
## [6,] 0.1604853
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] -1.9626047
## [2,] -0.2291284
## [3,] -0.2291284
## [4,] -0.2291284
## [5,] 0.3041491
## [6,] -0.2291284
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] -2.54378552
## [2,] -2.54378552
## [3,] -0.44374851
## [4,] -0.03466219
## [5,] -0.44838802
## [6,] -0.03466219
##
## Koefisien Jalur:
## Variabel Laten Perfomance Expectancy
## NULL
##
## Variabel Laten Effort Expectancy
## NULL
##
## Variabel Laten Social Influence
## NULL
##
## Variabel Laten Facilitating Conditions
## NULL
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.17984759
## [2,] -0.05737437
## [3,] 0.21005121
## [4,] 0.50697975
##
## Outer Loading:
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] 0.8715220
## [2,] 0.8914727
## [3,] 0.8319485
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] 0.8868386
## [2,] 0.8781684
## [3,] 0.8338407
##
## Variabel Laten Social Influence
## [,1]
## [1,] 0.9304480
## [2,] 0.7581066
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] 0.8037971
## [2,] 0.9068410
## [3,] 0.7269303
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.9253825
## [2,] 0.8702525
## [3,] 0.9257141
## [4,] 0.7465068
##
## Matriks Residual Gabungan:
## [,1] [,2] [,3] [,4] [,5] [,6]
## [1,] -0.08791535 -0.1296811 0.2475975 0.2664125 0.04963479 -0.3805315
## [2,] -0.08791535 -0.1296811 0.2475975 0.1304387 -0.01545051 -0.1523294
## [3,] -0.08791535 -0.1296811 0.2475975 0.7334609 0.58167623 -1.4280850
## [4,] 0.28525906 0.2520359 -0.6138829 0.2664125 0.04963479 -0.3805315
## [5,] -0.08791535 -0.1296811 0.2475975 -0.3366097 -0.54749195 0.8952241
## [6,] -0.08791535 -0.1296811 0.2475975 0.2664125 0.04963479 -0.3805315
## [,7] [,8] [,9] [,10] [,11] [,12]
## [1,] -0.29818677 0.53068795 -1.1022605 0.3221855 0.8362000 -0.1498453
## [2,] 0.50788054 -0.90388342 0.1732348 0.1394588 -0.4239163 -0.1498453
## [3,] 0.42197524 -0.75099634 0.1732348 0.1394588 -0.4239163 -0.8274396
## [4,] -0.02401106 0.04273288 0.1732348 0.1394588 -0.4239163 0.0597423
## [5,] -0.38409207 0.68357502 -0.2554121 -0.3441392 0.8381570 0.4425970
## [6,] -0.02401106 0.04273288 0.1732348 0.1394588 -0.4239163 0.0597423
## [,13] [,14] [,15] [,16]
## [1,] 0.5135670 -0.51187444 0.1696938 -1.12838771
## [2,] 0.5135670 -0.51187444 0.1696938 -1.97301066
## [3,] 0.1230986 0.30880705 0.6908102 -0.19022938
## [4,] -0.2329097 -0.06988993 0.3854245 0.12310332
## [5,] 0.1271362 0.31310191 -1.3945346 -0.42141872
## [6,] -0.2329097 -0.06988993 0.3854245 0.04609492
##
## CR:
## Variabel Laten Perfomance Expectancy
## [1] 0.8993515
##
## Variabel Laten Effort Expectancy
## [1] 0.9004078
##
## Variabel Laten Social Influence
## [1] 0.8359482
##
## Variabel Laten Facilitating Conditions
## [1] 0.8555591
##
## Variabel Laten Behavioral Intention
## [1] 0.9252116
##
## AVE:
## Variabel Laten Perfomance Expectancy
## [1] 0.7488042
##
## Variabel Laten Effort Expectancy
## [1] 0.7509843
##
## Variabel Laten Social Influence
## [1] 0.7202296
##
## Variabel Laten Facilitating Conditions
## [1] 0.665626
##
## Variabel Laten Behavioral Intention
## [1] 0.7569728
##
## HTMT Matrix:
## Perfomance Expectancy Effort Expectancy
## Perfomance Expectancy NA 0.3737800
## Effort Expectancy 0.3737800 NA
## Social Influence 0.6172193 0.6160426
## Facilitating Conditions 0.4824333 0.7979404
## Behavioral Intention 0.5167693 0.4471064
## Social Influence Facilitating Conditions
## Perfomance Expectancy 0.6172193 0.4824333
## Effort Expectancy 0.6160426 0.7979404
## Social Influence NA 0.8150128
## Facilitating Conditions 0.8150128 NA
## Behavioral Intention 0.6839490 0.7799284
## Behavioral Intention
## Perfomance Expectancy 0.5167693
## Effort Expectancy 0.4471064
## Social Influence 0.6839490
## Facilitating Conditions 0.7799284
## Behavioral Intention NA
##
## R Square:
## Variabel Laten Perfomance Expectancy
## [1] NA
##
## Variabel Laten Effort Expectancy
## [1] NA
##
## Variabel Laten Social Influence
## [1] NA
##
## Variabel Laten Facilitating Conditions
## [1] NA
##
## Variabel Laten Behavioral Intention
## [1] 0.5115431resultbootstrap <- plsalgorithm_bootstrap(data = inputdata$Data,inner_weight_scheme = "Path",matrix_outer_model = inputdata$MOM, matrix_inner_model = inputdata$MIM, indicator = inputdata$Indikator, laten_variables = inputdata$Variabel, n_bootstrap=50)
print(resultbootstrap$path_coefficient)## [[1]]
## data frame with 0 columns and 0 rows
##
## [[2]]
## data frame with 0 columns and 0 rows
##
## [[3]]
## data frame with 0 columns and 0 rows
##
## [[4]]
## data frame with 0 columns and 0 rows
##
## [[5]]
## path_coefficient_original average_path_coefficient standar_error t_hitung
## 1 0.17984759 0.16642143 0.06576088 2.734872
## 2 -0.05737437 -0.04801611 0.04064532 -1.411586
## 3 0.21005121 0.20546637 0.06606206 3.179604
## 4 0.50697975 0.51153925 0.07514495 6.746691
## p_value
## 1 6.240454e-03
## 2 1.580719e-01
## 3 1.474763e-03
## 4 1.512546e-11dfmatrikserror <- as.data.frame(hasil$`Matrix Residual`)
print(head(dfmatrikserror))## V1 V2 V3 V4 V5 V6
## 1 -0.08791535 -0.1296811 0.2475975 0.2664125 0.04963479 -0.3805315
## 2 -0.08791535 -0.1296811 0.2475975 0.1304387 -0.01545051 -0.1523294
## 3 -0.08791535 -0.1296811 0.2475975 0.7334609 0.58167623 -1.4280850
## 4 0.28525906 0.2520359 -0.6138829 0.2664125 0.04963479 -0.3805315
## 5 -0.08791535 -0.1296811 0.2475975 -0.3366097 -0.54749195 0.8952241
## 6 -0.08791535 -0.1296811 0.2475975 0.2664125 0.04963479 -0.3805315
## V7 V8 V9 V10 V11 V12
## 1 -0.29818677 0.53068795 -1.1022605 0.3221855 0.8362000 -0.1498453
## 2 0.50788054 -0.90388342 0.1732348 0.1394588 -0.4239163 -0.1498453
## 3 0.42197524 -0.75099634 0.1732348 0.1394588 -0.4239163 -0.8274396
## 4 -0.02401106 0.04273288 0.1732348 0.1394588 -0.4239163 0.0597423
## 5 -0.38409207 0.68357502 -0.2554121 -0.3441392 0.8381570 0.4425970
## 6 -0.02401106 0.04273288 0.1732348 0.1394588 -0.4239163 0.0597423
## V13 V14 V15 V16
## 1 0.5135670 -0.51187444 0.1696938 -1.12838771
## 2 0.5135670 -0.51187444 0.1696938 -1.97301066
## 3 0.1230986 0.30880705 0.6908102 -0.19022938
## 4 -0.2329097 -0.06988993 0.3854245 0.12310332
## 5 0.1271362 0.31310191 -1.3945346 -0.42141872
## 6 -0.2329097 -0.06988993 0.3854245 0.04609492# Jalankan algoritma Kernel K-Means
hasil_kkmeans <- kernel_kmeans(dfmatrikserror, k=2, max_iter = 100, kernel_fun = rbf_kernel, sigma = 1)## Kernel K-Means konvergen pada iterasi ke- 5# Tampilkan hasil cluster
print(hasil_kkmeans$clusters)## [1] 2 2 2 1 2 1 2 1 1 2 1 1 1 1 2 2 1 2 2 1 2 2 1 2 1 2 2 2 2 1 2 2 1 2 1 2 2
## [38] 2 2 2 1 1 1 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 2 2 2 2 2
## [75] 2 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 2 2 2 2 2 2 1 1 2
## [112] 1 2 1 2 1 1 2 1 2 2 2 1 2 1 2 2 2 2 1 2 2 2 2 1 2 2 2 2 2 1 2 2 2 2 1 2 2
## [149] 1 1 2 2 2 2 1 2 1 1 2 2 1 1 1 2 1 2 2 2 1 1 2 2 2 2 2 2 2 1 1 2 2 1 1 2 1
## [186] 2 1 1 2 1 1 1 1 2 2 1 2 2 1 2 2 1 2 1 2 2 2 2 2 1 1 1 2 1 2 2 2 2 2 2 2 2
## [223] 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 2 1 1 2 1 2 2 2 1 1 2 2 2 2 2 2 2 1 1 2 2
## [260] 1 1 2 2 1 1 2 1 1 1 2 2 1 2 2 2 2 2 2 1 2 1 1 1 2 1 2 2 2 1 2 1 2 1 1 2 1
## [297] 1 1 1 2 2 1 2 2 1 2 1 2 1 1 1 1 2 2 1 2 2 1 2 2 1 2 1 2 2 2 2 2 1 1 1 2 1
## [334] 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 2 2 2 2 2 2 1 1 2data_matrix=as.matrix(dfmatrikserror)
print(table(hasil_kkmeans$clusters))##
## 1 2
## 121 245data_terkelompok_kernel <- split(inputdata$Data,hasil_kkmeans$clusters)
hasil_k1=plsalgorithm(data = data_terkelompok_kernel$"1",inner_weight_scheme = "Path",matrix_outer_model = inputdata$MOM, matrix_inner_model = inputdata$MIM, indicator = inputdata$Indikator, laten_variables = inputdata$Variabel)## Number of Iterasi: 5
## Outer Weight
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] 0.3837338
## [2,] 0.3243786
## [3,] 0.3408206
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] 0.3333333
## [2,] 0.3333333
## [3,] 0.3333333
##
## Variabel Laten Social Influence
## [,1]
## [1,] 0.5319346
## [2,] 0.5102544
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] 0.3428930
## [2,] 0.3295562
## [3,] 0.3484066
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.2761567
## [2,] 0.2761567
## [3,] 0.2761567
## [4,] 0.2112467
##
## Skor Variabel Laten
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] -1.0248007
## [2,] -0.4894076
## [3,] -1.6642881
## [4,] -0.4894076
## [5,] -0.4894076
## [6,] 1.4285703
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] -0.8075271
## [2,] -0.8075271
## [3,] -0.8075271
## [4,] -0.8075271
## [5,] -0.8075271
## [6,] 1.2281142
##
## Variabel Laten Social Influence
## [,1]
## [1,] -0.4319952
## [2,] -0.4319952
## [3,] -0.4319952
## [4,] -0.4319952
## [5,] -0.4319952
## [6,] 1.4348411
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] -0.4990059
## [2,] -0.4990059
## [3,] -0.4990059
## [4,] -0.4990059
## [5,] -0.4990059
## [6,] 1.3538611
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] -0.4479663
## [2,] -0.4479663
## [3,] -0.4479663
## [4,] -0.4479663
## [5,] -0.4479663
## [6,] 1.1563593
##
## Koefisien Jalur:
## Variabel Laten Perfomance Expectancy
## NULL
##
## Variabel Laten Effort Expectancy
## NULL
##
## Variabel Laten Social Influence
## NULL
##
## Variabel Laten Facilitating Conditions
## NULL
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.01067045
## [2,] 0.03637059
## [3,] 0.20403669
## [4,] 0.76202446
##
## Outer Loading:
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] 0.9584553
## [2,] 0.9528553
## [3,] 0.9480721
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] 1
## [2,] 1
## [3,] 1
##
## Variabel Laten Social Influence
## [,1]
## [1,] 0.9612035
## [2,] 0.9577626
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] 0.9852916
## [2,] 0.9621544
## [3,] 0.9904132
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.9908249
## [2,] 0.9908249
## [3,] 0.9908249
## [4,] 0.8479716
##
## Matriks Residual Gabungan:
## [,1] [,2] [,3] [,4] [,5]
## [1,] 0.45409150 0.4062404 -0.89790921 2.220446e-16 2.220446e-16
## [2,] -0.05905885 -0.1039118 0.16539388 2.220446e-16 2.220446e-16
## [3,] 1.06701165 -0.9558438 -0.29162899 2.220446e-16 2.220446e-16
## [4,] -0.05905885 -0.1039118 0.16539388 2.220446e-16 2.220446e-16
## [5,] -0.05905885 -0.1039118 0.16539388 2.220446e-16 2.220446e-16
## [6,] 0.03913683 0.0399560 -0.08209303 -2.220446e-16 -2.220446e-16
## [,6] [,7] [,8] [,9] [,10] [,11]
## [1,] 2.220446e-16 -0.01009154 0.01052032 -0.03279443 0.09586926 -0.05840686
## [2,] 2.220446e-16 -0.01009154 0.01052032 -0.03279443 0.09586926 -0.05840686
## [3,] 2.220446e-16 -0.01009154 0.01052032 -0.03279443 0.09586926 -0.05840686
## [4,] 2.220446e-16 -0.01009154 0.01052032 -0.03279443 0.09586926 -0.05840686
## [5,] 2.220446e-16 -0.01009154 0.01052032 -0.03279443 0.09586926 -0.05840686
## [6,] -2.220446e-16 0.03351833 -0.03494249 0.00805459 -0.02635957 0.01700627
## [,12] [,13] [,14] [,15] [,16]
## [1,] -0.08427797 -0.08427797 -0.08427797 0.3305225 0.06073662
## [2,] -0.08427797 -0.08427797 -0.08427797 0.3305225 0.05502373
## [3,] -0.08427797 -0.08427797 -0.08427797 0.3305225 0.06756024
## [4,] -0.08427797 -0.08427797 -0.08427797 0.3305225 0.05502373
## [5,] -0.08427797 -0.08427797 -0.08427797 0.3305225 0.05502373
## [6,] 0.26260809 0.26260809 0.26260809 -1.0299000 -0.22798694
##
## CR:
## Variabel Laten Perfomance Expectancy
## [1] 0.9675067
##
## Variabel Laten Effort Expectancy
## [1] 1
##
## Variabel Laten Social Influence
## [1] 0.9586644
##
## Variabel Laten Facilitating Conditions
## [1] 0.986001
##
## Variabel Laten Behavioral Intention
## [1] 0.9775146
##
## AVE:
## Variabel Laten Perfomance Expectancy
## [1] 0.9084701
##
## Variabel Laten Effort Expectancy
## [1] 1
##
## Variabel Laten Social Influence
## [1] 0.9206107
##
## Variabel Laten Facilitating Conditions
## [1] 0.959153
##
## Variabel Laten Behavioral Intention
## [1] 0.9160644
##
## HTMT Matrix:
## Perfomance Expectancy Effort Expectancy
## Perfomance Expectancy NA 0.6923501
## Effort Expectancy 0.6923501 NA
## Social Influence 0.6849658 0.6076648
## Facilitating Conditions 0.8357941 0.7992708
## Behavioral Intention 0.8034515 0.7705566
## Social Influence Facilitating Conditions
## Perfomance Expectancy 0.6849658 0.8357941
## Effort Expectancy 0.6076648 0.7992708
## Social Influence NA 0.8360970
## Facilitating Conditions 0.8360970 NA
## Behavioral Intention 0.8878069 0.9844447
## Behavioral Intention
## Perfomance Expectancy 0.8034515
## Effort Expectancy 0.7705566
## Social Influence 0.8878069
## Facilitating Conditions 0.9844447
## Behavioral Intention NA
##
## R Square:
## Variabel Laten Perfomance Expectancy
## [1] NA
##
## Variabel Laten Effort Expectancy
## [1] NA
##
## Variabel Laten Social Influence
## [1] NA
##
## Variabel Laten Facilitating Conditions
## [1] NA
##
## Variabel Laten Behavioral Intention
## [1] 0.9388645resultbootstrap_k1 <- plsalgorithm_bootstrap(data = data_terkelompok_kernel$"1",inner_weight_scheme = "Path",matrix_outer_model = inputdata$MOM, matrix_inner_model = inputdata$MIM, indicator = inputdata$Indikator, laten_variables = inputdata$Variabel, n_bootstrap=500)
hasil_k2=plsalgorithm(data = data_terkelompok_kernel$"2",inner_weight_scheme = "Path",matrix_outer_model = inputdata$MOM, matrix_inner_model = inputdata$MIM, indicator = inputdata$Indikator, laten_variables = inputdata$Variabel)## Number of Iterasi: 9
## Outer Weight
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] 0.4240390
## [2,] 0.4527869
## [3,] 0.3042690
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] 0.5053360
## [2,] 0.2482742
## [3,] 0.4409510
##
## Variabel Laten Social Influence
## [,1]
## [1,] 0.8631433
## [2,] 0.2887810
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] 0.3768780
## [2,] 0.5697645
## [3,] 0.3177982
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.3509757
## [2,] 0.3486298
## [3,] 0.3012877
## [4,] 0.1576645
##
## Skor Variabel Laten
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] -0.10118
## [2,] -0.10118
## [3,] -0.10118
## [4,] -0.10118
## [5,] -0.10118
## [6,] -0.10118
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] -0.5927826
## [2,] 1.3248519
## [3,] 0.5388844
## [4,] 0.1931850
## [5,] -0.9690724
## [6,] -0.5927826
##
## Variabel Laten Social Influence
## [,1]
## [1,] -1.77242853
## [2,] -1.17860849
## [3,] 0.04094556
## [4,] -0.55287447
## [5,] 0.35381257
## [6,] 0.35381257
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] -1.7966877
## [2,] -0.1039619
## [3,] -0.1039619
## [4,] 0.3915052
## [5,] -0.5736815
## [6,] -0.1039619
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] -2.24776698
## [2,] -2.24776698
## [3,] -0.31207207
## [4,] -0.21203489
## [5,] 0.09640565
## [6,] 1.49596720
##
## Koefisien Jalur:
## Variabel Laten Perfomance Expectancy
## NULL
##
## Variabel Laten Effort Expectancy
## NULL
##
## Variabel Laten Social Influence
## NULL
##
## Variabel Laten Facilitating Conditions
## NULL
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.1775119
## [2,] -0.0751141
## [3,] 0.2174542
## [4,] 0.4574628
##
## Outer Loading:
## Variabel Laten Perfomance Expectancy
## [,1]
## [1,] 0.8522918
## [2,] 0.8923138
## [3,] 0.7709198
##
## Variabel Laten Effort Expectancy
## [,1]
## [1,] 0.8635967
## [2,] 0.8048812
## [3,] 0.8249495
##
## Variabel Laten Social Influence
## [,1]
## [1,] 0.9625411
## [2,] 0.5858731
##
## Variabel Laten Facilitating Conditions
## [,1]
## [1,] 0.7356745
## [2,] 0.9020437
## [3,] 0.6569828
##
## Variabel Laten Behavioral Intention
## [,1]
## [1,] 0.9147023
## [2,] 0.8515168
## [3,] 0.9112096
## [4,] 0.6822166
##
## Matriks Residual Gabungan:
## [,1] [,2] [,3] [,4] [,5] [,6]
## [1,] -0.08131099 -0.1305332 0.307566 0.3837992 0.06882952 -0.4785932
## [2,] -0.08131099 -0.1305332 0.307566 0.2225381 0.04098366 -0.2781073
## [3,] -0.08131099 -0.1305332 0.307566 0.9012971 0.67359415 -1.4121613
## [4,] -0.08131099 -0.1305332 0.307566 -0.2949598 -0.56378098 0.6554607
## [5,] -0.08131099 -0.1305332 0.307566 0.7087619 -1.14392398 -0.1681731
## [6,] -0.08131099 -0.1305332 0.307566 0.3837992 0.06882952 -0.4785932
## [,7] [,8] [,9] [,10] [,11] [,12]
## [1,] -0.10330742 0.3087776 -1.0132096 0.3525524 0.5694953 -0.076867637
## [2,] 0.37556420 -1.1225313 0.2341828 0.1477415 -0.5425965 -0.076867637
## [3,] 0.25214108 -0.7536295 0.2341828 0.1477415 -0.5425965 -0.683617501
## [4,] -0.22673054 0.6776794 -0.1303198 -0.2991915 0.6909520 0.388712982
## [5,] -0.04900629 0.1464759 -0.6666003 0.5714491 -0.2339989 0.106581705
## [6,] -0.04900629 0.1464759 0.2341828 0.1477415 -0.5425965 -0.009765741
## [,13] [,14] [,15] [,16]
## [1,] 0.4279465131 -0.45597519 0.09617572 -1.0669928
## [2,] 0.4279465131 -0.45597519 0.09617572 -1.8264392
## [3,] 0.0846371074 0.31540639 0.73192149 -0.2149787
## [4,] -0.0005462354 0.22425155 -1.29263448 -0.2384375
## [5,] -0.2631885482 -0.05680242 0.45325120 0.2270752
## [6,] -0.1499713791 -0.06449365 0.47660157 1.4400222
##
## CR:
## Variabel Laten Perfomance Expectancy
## [1] 0.877539
##
## Variabel Laten Effort Expectancy
## [1] 0.8703872
##
## Variabel Laten Social Influence
## [1] 0.7665277
##
## Variabel Laten Facilitating Conditions
## [1] 0.8127103
##
## Variabel Laten Behavioral Intention
## [1] 0.908082
##
## AVE:
## Variabel Laten Perfomance Expectancy
## [1] 0.7056476
##
## Variabel Laten Effort Expectancy
## [1] 0.6913916
##
## Variabel Laten Social Influence
## [1] 0.6348664
##
## Variabel Laten Facilitating Conditions
## [1] 0.5955087
##
## Variabel Laten Behavioral Intention
## [1] 0.7143709
##
## HTMT Matrix:
## Perfomance Expectancy Effort Expectancy
## Perfomance Expectancy NA 0.2129749
## Effort Expectancy 0.2129749 NA
## Social Influence 0.5786560 0.6518371
## Facilitating Conditions 0.2853007 0.7734470
## Behavioral Intention 0.3922148 0.3211511
## Social Influence Facilitating Conditions
## Perfomance Expectancy 0.5786560 0.2853007
## Effort Expectancy 0.6518371 0.7734470
## Social Influence NA 0.8214224
## Facilitating Conditions 0.8214224 NA
## Behavioral Intention 0.5747911 0.6840497
## Behavioral Intention
## Perfomance Expectancy 0.3922148
## Effort Expectancy 0.3211511
## Social Influence 0.5747911
## Facilitating Conditions 0.6840497
## Behavioral Intention NA
##
## R Square:
## Variabel Laten Perfomance Expectancy
## [1] NA
##
## Variabel Laten Effort Expectancy
## [1] NA
##
## Variabel Laten Social Influence
## [1] NA
##
## Variabel Laten Facilitating Conditions
## [1] NA
##
## Variabel Laten Behavioral Intention
## [1] 0.4021558resultbootstrap_k2 <- plsalgorithm_bootstrap(data = data_terkelompok_kernel$"2",inner_weight_scheme = "Path",matrix_outer_model = inputdata$MOM, matrix_inner_model = inputdata$MIM, indicator = inputdata$Indikator, laten_variables = inputdata$Variabel, n_bootstrap=500)
resultbootstrap_k1$path_coefficient## [[1]]
## data frame with 0 columns and 0 rows
##
## [[2]]
## data frame with 0 columns and 0 rows
##
## [[3]]
## data frame with 0 columns and 0 rows
##
## [[4]]
## data frame with 0 columns and 0 rows
##
## [[5]]
## path_coefficient_original average_path_coefficient standar_error t_hitung
## 1 0.01067045 0.01413965 0.06739955 0.1583163
## 2 0.03637059 0.03199431 0.02004987 1.8140064
## 3 0.20403669 0.20368414 0.08712341 2.3419273
## 4 0.76202446 0.76199453 0.11381854 6.6950822
## p_value
## 1 8.742075e-01
## 2 6.967675e-02
## 3 1.918445e-02
## 4 2.155498e-11resultbootstrap_k2$path_coefficient## [[1]]
## data frame with 0 columns and 0 rows
##
## [[2]]
## data frame with 0 columns and 0 rows
##
## [[3]]
## data frame with 0 columns and 0 rows
##
## [[4]]
## data frame with 0 columns and 0 rows
##
## [[5]]
## path_coefficient_original average_path_coefficient standar_error t_hitung
## 1 0.1775119 0.18588040 0.07788788 2.279069
## 2 -0.0751141 -0.07794092 0.05960726 -1.260150
## 3 0.2174542 0.22380002 0.08626445 2.520786
## 4 0.4574628 0.45371741 0.07991712 5.724215
## p_value
## 1 2.266294e-02
## 2 2.076152e-01
## 3 1.170930e-02
## 4 1.039130e-08