R/evaluation.R
In sanmai-NL/MScAnalysis: Exploiting Linguistic Structure with Trail N-grams in Linguistic Networks to Classify Dutch Legal Texts
Defines functions
evaluation_all_models
evaluate_results
statistical_analysis
#' @export
evaluation_all_models <- function(OUTPUT_DIR_PATH) {
TEMP_DIR_PATH <- tempdir()
setwd(TEMP_DIR_PATH)
futile.logger::flog.info("Writing evaluation report to '%s' ...", TEMP_DIR_PATH)
MODELS_REL_FILES_PATHS_CVEC <-
base::list.files(
OUTPUT_DIR_PATH,
pattern='^model\\.rds$',
recursive=TRUE)
MODELS_LST <-
parallel::mclapply(
base::file.path(OUTPUT_DIR_PATH, MODELS_REL_FILES_PATHS_CVEC),
base::readRDS)
stopifnot(length(MODELS_REL_FILES_PATHS_CVEC) > 0)
###
## Collect classification measure scores
###
MODELS_K_SCORES_LST <-
base::vapply(
MODELS_LST,
function(MODEL) K_SLC(get_confusion_matrix_of_model(MODEL)),
FUN.VALUE=base::numeric(length=1L))
RESULTS_CMAT <-
stringi::stri_split_fixed(
str=MODELS_REL_FILES_PATHS_CVEC,
pattern='/',
simplify=TRUE)
CONDITION1_CMAT <-
stringi::stri_split_fixed(
RESULTS_CMAT[, 1L],
pattern=base::c(' '),
simplify=TRUE)
CONDITION3_CMAT <-
base::matrix(
stringi::stri_split_fixed(
CONDITION1_CMAT,
pattern=base::c('='),
simplify=TRUE),
nrow=nrow(CONDITION1_CMAT))
CONDITION2_CMAT <-
stringi::stri_split_fixed(
RESULTS_CMAT[, 2L],
pattern='.',
simplify=TRUE)
FINAL_RESULTS_CMAT <- base::cbind(
CONDITION3_CMAT[, 1L],
CONDITION2_CMAT[, base::c(1L, 2L, 5L)],
RESULTS_CMAT[, base::c(-1L, -2L, -4L)],
CONDITION3_CMAT[, 9L:14L]
)
base::colnames(FINAL_RESULTS_CMAT) <-
base::c('subset', 'feature_set', 'task', 'learner', 'run', 'k', 'feature_scoring',
'SRF_N', 'STRCT_I', 'STRCT_N', 'STRCT_M')
RESULTS_D_F <-
dplyr::mutate(
dplyr::as_data_frame(
base::cbind(
base::as.data.frame(FINAL_RESULTS_CMAT),
K_score=MODELS_K_SCORES_LST,
file_path=MODELS_REL_FILES_PATHS_CVEC)))
SRF_BODY_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'body:',
feature_set == 'SRF')
STRCT_BODY_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'body:',
feature_set == 'STRCT')
SRF_AND_STRCT_BODY_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'body:',
feature_set == 'SRF&STRCT')
SRF_CONCLUSION_BY_TYPE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_by_type:',
feature_set == 'SRF')
STRCT_CONCLUSION_BY_TYPE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_by_type:',
feature_set == 'STRCT')
SRF_AND_STRCT_CONCLUSION_BY_TYPE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_by_type:',
feature_set == 'SRF&STRCT')
SRF_CONCLUSION_WHOLE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_whole:',
feature_set == 'SRF')
STRCT_CONCLUSION_WHOLE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_whole:',
feature_set == 'STRCT')
SRF_AND_STRCT_CONCLUSION_WHOLE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_whole:',
feature_set == 'SRF&STRCT')
##
SUBSETS_CVEC <-
c('body', 'conclusion_by_type',
'conclusion_whole', 'conclusion_whole', 'conclusion_whole')
TASKS_CVEC <-
c('conclusion_type_I',
'conclusion_type_I',
'conclusion_type_is_denied_I',
'conclusion_type_is_held_insufficient_I',
'conclusion_type_is_granted_I')
TASKS_LST <-
base::list(
base::list(
SRF=SRF_BODY_D_F,
STRCT=STRCT_BODY_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_BODY_D_F),
base::list(
SRF=SRF_CONCLUSION_BY_TYPE_D_F,
STRCT=STRCT_CONCLUSION_BY_TYPE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_BY_TYPE_D_F),
base::list(
SRF=SRF_CONCLUSION_WHOLE_D_F,
STRCT=STRCT_CONCLUSION_WHOLE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_WHOLE_D_F),
base::list(
SRF=SRF_CONCLUSION_WHOLE_D_F,
STRCT=STRCT_CONCLUSION_WHOLE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_WHOLE_D_F),
base::list(
SRF=SRF_CONCLUSION_WHOLE_D_F,
STRCT=STRCT_CONCLUSION_WHOLE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_WHOLE_D_F))
EFFECTIVENESS_OVERVIEW_STR <- 'Subset\tTask\tFeature set\tM\tSD\n'
for (TASK_LST_INDEX_I in seq_along(TASKS_LST)) {
TASK_LST <- TASKS_LST[[TASK_LST_INDEX_I]]
TASK_STR <- TASKS_CVEC[TASK_LST_INDEX_I]
SUBSET_STR <- SUBSETS_CVEC[TASK_LST_INDEX_I]
# TODO: (efficiency)
FEATURE_SET_LST <-
base::lapply(
TASK_LST,
function(D_F) dplyr::filter(D_F, task == TASK_STR)[['K_score']])
names(FEATURE_SET_LST) <- NULL
SCORE_BEANPLOT_SVG_FILE_PATH_STR <-
stringi::stri_join(TASK_STR, SUBSET_STR, 'K_score', 'svg', sep='.')
svg(file=SCORE_BEANPLOT_SVG_FILE_PATH_STR, antialias='none', width=15, height=15)
base::do.call(
beanplot::beanplot,
base::list(FEATURE_SET_LST,
col=c('#8dd3c7','#ffffb3','#bebada'),
main=stringi::stri_join("Classifiers' K scores (per CV run)"),
names=c('SRF', 'STRCT', 'SRF & STRCT'),
what=c(1, 1, 1, 1),
ylab='K score',
add=FALSE))
dev.off()
for (FEATURE_REPRESENTATION_STR in names(TASK_LST)) {
TASK_FEATURE_SET_D_F <-
dplyr::filter(TASK_LST[[FEATURE_REPRESENTATION_STR]],
task == TASK_STR)
# TODO: do not average across runs?
CONFIGURATION_RANKING_D_F <-
dplyr::arrange(
dplyr::ungroup(
dplyr::summarize(
dplyr::group_by(
dplyr::select(
TASK_FEATURE_SET_D_F,
run, learner, k, feature_scoring,
SRF_N, STRCT_I, STRCT_N, STRCT_M,
K_score),
# ~setdiff(colnames(TASK_FEATURE_SET_D_F), 'run')),
learner, k, feature_scoring, SRF_N, STRCT_I, STRCT_N, STRCT_M),
mean_K_score=mean(K_score))),
desc(mean_K_score))
utils::write.csv(
CONFIGURATION_RANKING_D_F,
file=stringi::stri_join(TASK_STR, SUBSET_STR, FEATURE_REPRESENTATION_STR,
'K_score', 'csv', sep='.'),
row.names=FALSE)
EFFECTIVENESS_OVERVIEW_STR <-
stringi::stri_join(
EFFECTIVENESS_OVERVIEW_STR,
sprintf('%s\t%s\t%s\t%.2f\t%.2f\n',
SUBSET_STR,
TASK_STR, FEATURE_REPRESENTATION_STR,
mean(CONFIGURATION_RANKING_D_F$mean_K_score),
sd(CONFIGURATION_RANKING_D_F$mean_K_score)))
# message(EFFECTIVENESS_OVERVIEW_STR)
}
}
cat(file='effectiveness_overview.tsv', EFFECTIVENESS_OVERVIEW_STR)
utils::write.csv(
RESULTS_D_F,
file='results.csv',
row.names=FALSE)
statistical_analysis(
RESULTS_DF=data.frame(RESULTS_D_F),
OUTPUT_DIR_PATH=TEMP_DIR_PATH)
REPORT_DIR_PATH_STR <- file.path(OUTPUT_DIR_PATH, 'report')
futile.logger::flog.info(
"Moving evaluation report to '%s' ...",
REPORT_DIR_PATH_STR)
setwd(OUTPUT_DIR_PATH)
move_or_copy(
from=TEMP_DIR_PATH,
to=REPORT_DIR_PATH_STR,
to_base=OUTPUT_DIR_PATH,
recursive=TRUE)
}
evaluate_results <- function(RESULTS_DF) {
RESULTS_DF$run <- as.factor(RESULTS_DF$run)
RESULTS_DF$SRF_N <- as.factor(RESULTS_DF$SRF_N)
RESULTS_DF$STRCT_N <- as.factor(RESULTS_DF$STRCT_N)
RESULTS_DF$k <- as.factor(RESULTS_DF$k)
RESULTS_DF$STRCT_M <- NULL
RESULTS_DF$STRCT_I <- NULL
RESULTS_DF$file_path <- NULL
STRCTK <- subset(RESULTS_DF, feature_set=='STRCT')$K_score
SRFK <- subset(RESULTS_DF, feature_set=='SRF')$K_score
STRCTANDSRFK <- subset(RESULTS_DF, feature_set=='SRF&STRCT')$K_score
return(RESULTS_DF)
}
statistical_analysis <- function(RESULTS_DF, OUTPUT_DIR_PATH) {
## Estimate linear regression model of configurations on K score
CLEANED_UP_RESULTS_DF <- get_data(RESULTS_DF)
BF_ALL_F <- formula(K_score ~.)
BF_SPECIFIC_F <-
formula(
K_score ~ subset * task + subset * learner + feature_set * learner +
task + learner + run + k + feature_scoring + SRF_N + STRCT_N + feature_set * SRF_N + feature_set * STRCT_N)
# RESULTS_LM <- glm(BF_SPECIFIC_F, data=CLEANED_UP_RESULTS_DF)
RESULTS_LM <- lm(BF_SPECIFIC_F, data=CLEANED_UP_RESULTS_DF)
svg(file.path(OUTPUT_DIR_PATH, 'regression_residuals_dist.svg'))
qqPlot(RESULTS_LM, main="QQ Plot")
#@ Plot distribution of studentized residuals
library(MASS)
sresid <- studres(RESULTS_LM)
hist(sresid, freq=FALSE, main="Distribution of Studentized residuals", xlab='Residuals of regression model')
xfit <- seq(min(sresid), max(sresid), length=40)
yfit <- dnorm(xfit)
lines(xfit, yfit)
dev.off()
## Produce regression model table
library(xtable)
idx <- order(abs(coef(summary(RESULTS_LM))[,1L]), decreasing=TRUE)
RESULTS_LM_TABLE <- xtable(coef(summary(RESULTS_LM))[idx,])
RESULTS_LM_TABLE_STR <- capture.output(print(RESULTS_LM_TABLE, type='html'))
cat(file=base::file.path(OUTPUT_DIR_PATH, 'results_lm_table.html'), RESULTS_LM_TABLE_STR)
## Test hypothesis
binom.test(
length(which(STRCTANDSRFK < SRFK)),
length(SRFK),
p=0.5,
alternative=base::c("greater"),
conf.level=0.95)
}
sanmai-NL/MScAnalysis documentation built on May 29, 2019, 1:47 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.
Defines functions evaluation_all_models evaluate_results statistical_analysis
#' @export
evaluation_all_models <- function(OUTPUT_DIR_PATH) {
TEMP_DIR_PATH <- tempdir()
setwd(TEMP_DIR_PATH)
futile.logger::flog.info("Writing evaluation report to '%s' ...", TEMP_DIR_PATH)
MODELS_REL_FILES_PATHS_CVEC <-
base::list.files(
OUTPUT_DIR_PATH,
pattern='^model\\.rds$',
recursive=TRUE)
MODELS_LST <-
parallel::mclapply(
base::file.path(OUTPUT_DIR_PATH, MODELS_REL_FILES_PATHS_CVEC),
base::readRDS)
stopifnot(length(MODELS_REL_FILES_PATHS_CVEC) > 0)
###
## Collect classification measure scores
###
MODELS_K_SCORES_LST <-
base::vapply(
MODELS_LST,
function(MODEL) K_SLC(get_confusion_matrix_of_model(MODEL)),
FUN.VALUE=base::numeric(length=1L))
RESULTS_CMAT <-
stringi::stri_split_fixed(
str=MODELS_REL_FILES_PATHS_CVEC,
pattern='/',
simplify=TRUE)
CONDITION1_CMAT <-
stringi::stri_split_fixed(
RESULTS_CMAT[, 1L],
pattern=base::c(' '),
simplify=TRUE)
CONDITION3_CMAT <-
base::matrix(
stringi::stri_split_fixed(
CONDITION1_CMAT,
pattern=base::c('='),
simplify=TRUE),
nrow=nrow(CONDITION1_CMAT))
CONDITION2_CMAT <-
stringi::stri_split_fixed(
RESULTS_CMAT[, 2L],
pattern='.',
simplify=TRUE)
FINAL_RESULTS_CMAT <- base::cbind(
CONDITION3_CMAT[, 1L],
CONDITION2_CMAT[, base::c(1L, 2L, 5L)],
RESULTS_CMAT[, base::c(-1L, -2L, -4L)],
CONDITION3_CMAT[, 9L:14L]
)
base::colnames(FINAL_RESULTS_CMAT) <-
base::c('subset', 'feature_set', 'task', 'learner', 'run', 'k', 'feature_scoring',
'SRF_N', 'STRCT_I', 'STRCT_N', 'STRCT_M')
RESULTS_D_F <-
dplyr::mutate(
dplyr::as_data_frame(
base::cbind(
base::as.data.frame(FINAL_RESULTS_CMAT),
K_score=MODELS_K_SCORES_LST,
file_path=MODELS_REL_FILES_PATHS_CVEC)))
SRF_BODY_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'body:',
feature_set == 'SRF')
STRCT_BODY_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'body:',
feature_set == 'STRCT')
SRF_AND_STRCT_BODY_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'body:',
feature_set == 'SRF&STRCT')
SRF_CONCLUSION_BY_TYPE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_by_type:',
feature_set == 'SRF')
STRCT_CONCLUSION_BY_TYPE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_by_type:',
feature_set == 'STRCT')
SRF_AND_STRCT_CONCLUSION_BY_TYPE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_by_type:',
feature_set == 'SRF&STRCT')
SRF_CONCLUSION_WHOLE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_whole:',
feature_set == 'SRF')
STRCT_CONCLUSION_WHOLE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_whole:',
feature_set == 'STRCT')
SRF_AND_STRCT_CONCLUSION_WHOLE_D_F <-
dplyr::filter(
RESULTS_D_F,
subset == 'conclusion_whole:',
feature_set == 'SRF&STRCT')
##
SUBSETS_CVEC <-
c('body', 'conclusion_by_type',
'conclusion_whole', 'conclusion_whole', 'conclusion_whole')
TASKS_CVEC <-
c('conclusion_type_I',
'conclusion_type_I',
'conclusion_type_is_denied_I',
'conclusion_type_is_held_insufficient_I',
'conclusion_type_is_granted_I')
TASKS_LST <-
base::list(
base::list(
SRF=SRF_BODY_D_F,
STRCT=STRCT_BODY_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_BODY_D_F),
base::list(
SRF=SRF_CONCLUSION_BY_TYPE_D_F,
STRCT=STRCT_CONCLUSION_BY_TYPE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_BY_TYPE_D_F),
base::list(
SRF=SRF_CONCLUSION_WHOLE_D_F,
STRCT=STRCT_CONCLUSION_WHOLE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_WHOLE_D_F),
base::list(
SRF=SRF_CONCLUSION_WHOLE_D_F,
STRCT=STRCT_CONCLUSION_WHOLE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_WHOLE_D_F),
base::list(
SRF=SRF_CONCLUSION_WHOLE_D_F,
STRCT=STRCT_CONCLUSION_WHOLE_D_F,
SRF_AND_STRCT=SRF_AND_STRCT_CONCLUSION_WHOLE_D_F))
EFFECTIVENESS_OVERVIEW_STR <- 'Subset\tTask\tFeature set\tM\tSD\n'
for (TASK_LST_INDEX_I in seq_along(TASKS_LST)) {
TASK_LST <- TASKS_LST[[TASK_LST_INDEX_I]]
TASK_STR <- TASKS_CVEC[TASK_LST_INDEX_I]
SUBSET_STR <- SUBSETS_CVEC[TASK_LST_INDEX_I]
# TODO: (efficiency)
FEATURE_SET_LST <-
base::lapply(
TASK_LST,
function(D_F) dplyr::filter(D_F, task == TASK_STR)[['K_score']])
names(FEATURE_SET_LST) <- NULL
SCORE_BEANPLOT_SVG_FILE_PATH_STR <-
stringi::stri_join(TASK_STR, SUBSET_STR, 'K_score', 'svg', sep='.')
svg(file=SCORE_BEANPLOT_SVG_FILE_PATH_STR, antialias='none', width=15, height=15)
base::do.call(
beanplot::beanplot,
base::list(FEATURE_SET_LST,
col=c('#8dd3c7','#ffffb3','#bebada'),
main=stringi::stri_join("Classifiers' K scores (per CV run)"),
names=c('SRF', 'STRCT', 'SRF & STRCT'),
what=c(1, 1, 1, 1),
ylab='K score',
add=FALSE))
dev.off()
for (FEATURE_REPRESENTATION_STR in names(TASK_LST)) {
TASK_FEATURE_SET_D_F <-
dplyr::filter(TASK_LST[[FEATURE_REPRESENTATION_STR]],
task == TASK_STR)
# TODO: do not average across runs?
CONFIGURATION_RANKING_D_F <-
dplyr::arrange(
dplyr::ungroup(
dplyr::summarize(
dplyr::group_by(
dplyr::select(
TASK_FEATURE_SET_D_F,
run, learner, k, feature_scoring,
SRF_N, STRCT_I, STRCT_N, STRCT_M,
K_score),
# ~setdiff(colnames(TASK_FEATURE_SET_D_F), 'run')),
learner, k, feature_scoring, SRF_N, STRCT_I, STRCT_N, STRCT_M),
mean_K_score=mean(K_score))),
desc(mean_K_score))
utils::write.csv(
CONFIGURATION_RANKING_D_F,
file=stringi::stri_join(TASK_STR, SUBSET_STR, FEATURE_REPRESENTATION_STR,
'K_score', 'csv', sep='.'),
row.names=FALSE)
EFFECTIVENESS_OVERVIEW_STR <-
stringi::stri_join(
EFFECTIVENESS_OVERVIEW_STR,
sprintf('%s\t%s\t%s\t%.2f\t%.2f\n',
SUBSET_STR,
TASK_STR, FEATURE_REPRESENTATION_STR,
mean(CONFIGURATION_RANKING_D_F$mean_K_score),
sd(CONFIGURATION_RANKING_D_F$mean_K_score)))
# message(EFFECTIVENESS_OVERVIEW_STR)
}
}
cat(file='effectiveness_overview.tsv', EFFECTIVENESS_OVERVIEW_STR)
utils::write.csv(
RESULTS_D_F,
file='results.csv',
row.names=FALSE)
statistical_analysis(
RESULTS_DF=data.frame(RESULTS_D_F),
OUTPUT_DIR_PATH=TEMP_DIR_PATH)
REPORT_DIR_PATH_STR <- file.path(OUTPUT_DIR_PATH, 'report')
futile.logger::flog.info(
"Moving evaluation report to '%s' ...",
REPORT_DIR_PATH_STR)
setwd(OUTPUT_DIR_PATH)
move_or_copy(
from=TEMP_DIR_PATH,
to=REPORT_DIR_PATH_STR,
to_base=OUTPUT_DIR_PATH,
recursive=TRUE)
}
evaluate_results <- function(RESULTS_DF) {
RESULTS_DF$run <- as.factor(RESULTS_DF$run)
RESULTS_DF$SRF_N <- as.factor(RESULTS_DF$SRF_N)
RESULTS_DF$STRCT_N <- as.factor(RESULTS_DF$STRCT_N)
RESULTS_DF$k <- as.factor(RESULTS_DF$k)
RESULTS_DF$STRCT_M <- NULL
RESULTS_DF$STRCT_I <- NULL
RESULTS_DF$file_path <- NULL
STRCTK <- subset(RESULTS_DF, feature_set=='STRCT')$K_score
SRFK <- subset(RESULTS_DF, feature_set=='SRF')$K_score
STRCTANDSRFK <- subset(RESULTS_DF, feature_set=='SRF&STRCT')$K_score
return(RESULTS_DF)
}
statistical_analysis <- function(RESULTS_DF, OUTPUT_DIR_PATH) {
## Estimate linear regression model of configurations on K score
CLEANED_UP_RESULTS_DF <- get_data(RESULTS_DF)
BF_ALL_F <- formula(K_score ~.)
BF_SPECIFIC_F <-
formula(
K_score ~ subset * task + subset * learner + feature_set * learner +
task + learner + run + k + feature_scoring + SRF_N + STRCT_N + feature_set * SRF_N + feature_set * STRCT_N)
# RESULTS_LM <- glm(BF_SPECIFIC_F, data=CLEANED_UP_RESULTS_DF)
RESULTS_LM <- lm(BF_SPECIFIC_F, data=CLEANED_UP_RESULTS_DF)
svg(file.path(OUTPUT_DIR_PATH, 'regression_residuals_dist.svg'))
qqPlot(RESULTS_LM, main="QQ Plot")
#@ Plot distribution of studentized residuals
library(MASS)
sresid <- studres(RESULTS_LM)
hist(sresid, freq=FALSE, main="Distribution of Studentized residuals", xlab='Residuals of regression model')
xfit <- seq(min(sresid), max(sresid), length=40)
yfit <- dnorm(xfit)
lines(xfit, yfit)
dev.off()
## Produce regression model table
library(xtable)
idx <- order(abs(coef(summary(RESULTS_LM))[,1L]), decreasing=TRUE)
RESULTS_LM_TABLE <- xtable(coef(summary(RESULTS_LM))[idx,])
RESULTS_LM_TABLE_STR <- capture.output(print(RESULTS_LM_TABLE, type='html'))
cat(file=base::file.path(OUTPUT_DIR_PATH, 'results_lm_table.html'), RESULTS_LM_TABLE_STR)
## Test hypothesis
binom.test(
length(which(STRCTANDSRFK < SRFK)),
length(SRFK),
p=0.5,
alternative=base::c("greater"),
conf.level=0.95)
}
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.