inst/extdata/test_df.md
In mbadge/AnalysisToolkitR: Objects and Functions for general data analysis
Test Dataset Classifiers
mabadgeley
r Sys.Date()
Test Dataset
data(test_df, package="ProjUtils")
str(test_df)
#> Classes 'tbl_df', 'tbl' and 'data.frame': 180 obs. of 14 variables:
#> $ id : chr "7862738" "7863668" "7872384" "7873551" ...
#> $ cnn_41 : num 0.992 1 1 1 1 ...
#> $ cnn_42 : num 0.8449 0.0296 0.8085 0.0559 0.9171 ...
#> $ cnn_43 : num 0.1781 0.0314 0.2015 0.5291 0.3774 ...
#> $ cnn_8 : num 0.25347 0.65845 0.40168 0.76503 0.00696 ...
#> $ cnn_10 : num 0.196 0.705 0.113 0.117 0.106 ...
#> $ nlp_4 : logi TRUE TRUE TRUE TRUE TRUE TRUE ...
#> $ nlp_8 : logi FALSE FALSE FALSE FALSE FALSE FALSE ...
#> $ nlp_10 : logi FALSE FALSE FALSE FALSE FALSE FALSE ...
#> $ attg : logi FALSE FALSE TRUE TRUE TRUE TRUE ...
#> $ javin : logi FALSE FALSE FALSE TRUE TRUE TRUE ...
#> $ eko : logi TRUE FALSE TRUE TRUE TRUE TRUE ...
#> $ mike : logi TRUE FALSE FALSE TRUE TRUE TRUE ...
#> $ priority: logi FALSE FALSE FALSE FALSE FALSE FALSE ...
Craft S4 Classifier lists
kGOLD_STD <- "attg"
# Make classifierPoints with logical columns
cPs <- test_df %>%
keep(.p=is.logical) %>%
`[`(names(.) %ni% kGOLD_STD) %>%
map(function(x_lgl) ClassifierPoint(Y_=x_lgl, Y=test_df[[kGOLD_STD]], id = test_df[["id"]]))
# Make classifierCurves with numeric columns
cCs <- test_df %>%
keep(.p=is.numeric) %>%
map(function(x_dbl) ClassifierCurve(pY=x_dbl, Y=test_df[[kGOLD_STD]], id = test_df[["id"]]))
Tabulate
glance() Binary Classifier Performance metrics
(lift_dl(rbind))(map(cPs, glance_ClassifierPoint)) %>%
tibble::rownames_to_column(var="grader") %>%
arrange(desc(acc)) %>%
knitr::kable(digits=2)
grader acc sens spec ppv npv fpr
javin 0.90 0.80 0.94 0.87 0.91 0.06
eko 0.80 0.82 0.79 0.64 0.91 0.21
mike 0.79 0.73 0.82 0.65 0.87 0.18
nlp_10 0.71 0.18 0.95 0.62 0.72 0.05
nlp_8 0.64 0.29 0.80 0.39 0.71 0.20
nlp_4 0.60 0.66 0.57 0.41 0.79 0.43
priority 0.55 0.16 0.73 0.21 0.66 0.27
glance() Continuous Classifier Performance metrics
(lift_dl(rbind))(map(cCs, glance_ClassifierCurve)) %>%
tibble::rownames_to_column(var="grader") %>%
arrange(desc(auc)) %>%
knitr::kable(digits=2)
grader auc acc sens spec ppv npv fpr
cnn_41 0.66 0.53 0.79 0.42 0.38 0.81 0.58
cnn_10 0.65 0.54 0.79 0.44 0.39 0.82 0.56
cnn_8 0.56 0.41 0.79 0.23 0.32 0.71 0.77
cnn_42 0.47 0.33 0.79 0.12 0.29 0.56 0.88
cnn_43 0.38 0.31 0.79 0.09 0.28 0.48 0.91
Pick a single cnn_4
map_dfr(cCs, roc, .id = "cnn") %>%
ggplot(., aes(x=x, y=y, col=cnn)) +
geom_line() +
AnalysisToolkit:::gg_roc_layers
cCs["cnn_4"] <- cCs["cnn_41"]
cCs[c("cnn_41", "cnn_42", "cnn_43")] <- NULL
Re-rank continuous classifiers (cnns)
(lift_dl(rbind))(map(cCs, glance_ClassifierCurve)) %>%
tibble::rownames_to_column() %>%
arrange(desc(auc)) %>%
knitr::kable(digits=2)
rowname auc acc sens spec ppv npv fpr
cnn_4 0.66 0.53 0.79 0.42 0.38 0.81 0.58
cnn_10 0.65 0.54 0.79 0.44 0.39 0.82 0.56
cnn_8 0.56 0.41 0.79 0.23 0.32 0.71 0.77
Save Classifier Lists As Package Data
classifiers <- list(
"classifier_points" = cPs,
"classifier_curves" = cCs
)
str(classifiers)
#> List of 2
#> $ classifier_points:List of 7
#> ..$ nlp_4 :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] TRUE TRUE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ nlp_8 :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE FALSE FALSE FALSE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ nlp_10 :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE FALSE FALSE FALSE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ javin :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ eko :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] TRUE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ mike :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] TRUE FALSE FALSE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ priority:Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE FALSE FALSE FALSE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> $ classifier_curves:List of 3
#> ..$ cnn_8 :Formal class 'ClassifierCurve' [package "AnalysisToolkit"] w..
#> .. .. ..@ pY: num [1:180] 0.25347 0.65845 0.40168 0.76503 0.00696 ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ cnn_10:Formal class 'ClassifierCurve' [package "AnalysisToolkit"] w..
#> .. .. ..@ pY: num [1:180] 0.196 0.705 0.113 0.117 0.106 ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ cnn_4 :Formal class 'ClassifierCurve' [package "AnalysisToolkit"] w..
#> .. .. ..@ pY: num [1:180] 0.992 1 1 1 1 ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
# devtools::use_data(classifiers)
Sys Info
Sys.info()
#> sysname release version nodename
#> "Windows" "10 x64" "build 16299" "MARCUSDESKTOP"
#> machine login user effective_user
#> "x86-64" "Marcus User" "Marcus User" "Marcus User"
search()
#> [1] ".GlobalEnv" "package:bindrcpp"
#> [3] "package:pander" "package:knitr"
#> [5] "package:vizR" "package:cowplot"
#> [7] "package:AnalysisToolkit" "package:broom"
#> [9] "package:ProjUtils" "package:MyUtils"
#> [11] "package:forcats" "package:purrr"
#> [13] "package:shiny" "package:magrittr"
#> [15] "package:rebus" "package:glue"
#> [17] "package:stringr" "package:tidyr"
#> [19] "package:dplyr" "package:plyr"
#> [21] "package:readr" "package:beepr"
#> [23] "package:ggplot2" "package:stats"
#> [25] "package:graphics" "package:grDevices"
#> [27] "package:utils" "package:datasets"
#> [29] "package:testthat" "package:assertive"
#> [31] "package:assertthat" "package:methods"
#> [33] "Autoloads" "package:base"
mbadge/AnalysisToolkitR documentation built on May 27, 2019, 1:08 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.
Test Dataset Classifiers
mabadgeley
r Sys.Date()
Test Dataset
data(test_df, package="ProjUtils")
str(test_df)
#> Classes 'tbl_df', 'tbl' and 'data.frame': 180 obs. of 14 variables:
#> $ id : chr "7862738" "7863668" "7872384" "7873551" ...
#> $ cnn_41 : num 0.992 1 1 1 1 ...
#> $ cnn_42 : num 0.8449 0.0296 0.8085 0.0559 0.9171 ...
#> $ cnn_43 : num 0.1781 0.0314 0.2015 0.5291 0.3774 ...
#> $ cnn_8 : num 0.25347 0.65845 0.40168 0.76503 0.00696 ...
#> $ cnn_10 : num 0.196 0.705 0.113 0.117 0.106 ...
#> $ nlp_4 : logi TRUE TRUE TRUE TRUE TRUE TRUE ...
#> $ nlp_8 : logi FALSE FALSE FALSE FALSE FALSE FALSE ...
#> $ nlp_10 : logi FALSE FALSE FALSE FALSE FALSE FALSE ...
#> $ attg : logi FALSE FALSE TRUE TRUE TRUE TRUE ...
#> $ javin : logi FALSE FALSE FALSE TRUE TRUE TRUE ...
#> $ eko : logi TRUE FALSE TRUE TRUE TRUE TRUE ...
#> $ mike : logi TRUE FALSE FALSE TRUE TRUE TRUE ...
#> $ priority: logi FALSE FALSE FALSE FALSE FALSE FALSE ...
Craft S4 Classifier lists
kGOLD_STD <- "attg"
# Make classifierPoints with logical columns
cPs <- test_df %>%
keep(.p=is.logical) %>%
`[`(names(.) %ni% kGOLD_STD) %>%
map(function(x_lgl) ClassifierPoint(Y_=x_lgl, Y=test_df[[kGOLD_STD]], id = test_df[["id"]]))
# Make classifierCurves with numeric columns
cCs <- test_df %>%
keep(.p=is.numeric) %>%
map(function(x_dbl) ClassifierCurve(pY=x_dbl, Y=test_df[[kGOLD_STD]], id = test_df[["id"]]))
Tabulate
glance() Binary Classifier Performance metrics
(lift_dl(rbind))(map(cPs, glance_ClassifierPoint)) %>%
tibble::rownames_to_column(var="grader") %>%
arrange(desc(acc)) %>%
knitr::kable(digits=2)
grader acc sens spec ppv npv fpr
javin 0.90 0.80 0.94 0.87 0.91 0.06 eko 0.80 0.82 0.79 0.64 0.91 0.21 mike 0.79 0.73 0.82 0.65 0.87 0.18 nlp_10 0.71 0.18 0.95 0.62 0.72 0.05 nlp_8 0.64 0.29 0.80 0.39 0.71 0.20 nlp_4 0.60 0.66 0.57 0.41 0.79 0.43 priority 0.55 0.16 0.73 0.21 0.66 0.27
glance() Continuous Classifier Performance metrics
(lift_dl(rbind))(map(cCs, glance_ClassifierCurve)) %>%
tibble::rownames_to_column(var="grader") %>%
arrange(desc(auc)) %>%
knitr::kable(digits=2)
grader auc acc sens spec ppv npv fpr
cnn_41 0.66 0.53 0.79 0.42 0.38 0.81 0.58 cnn_10 0.65 0.54 0.79 0.44 0.39 0.82 0.56 cnn_8 0.56 0.41 0.79 0.23 0.32 0.71 0.77 cnn_42 0.47 0.33 0.79 0.12 0.29 0.56 0.88 cnn_43 0.38 0.31 0.79 0.09 0.28 0.48 0.91
Pick a single cnn_4
map_dfr(cCs, roc, .id = "cnn") %>%
ggplot(., aes(x=x, y=y, col=cnn)) +
geom_line() +
AnalysisToolkit:::gg_roc_layers
cCs["cnn_4"] <- cCs["cnn_41"]
cCs[c("cnn_41", "cnn_42", "cnn_43")] <- NULL
Re-rank continuous classifiers (cnns)
(lift_dl(rbind))(map(cCs, glance_ClassifierCurve)) %>%
tibble::rownames_to_column() %>%
arrange(desc(auc)) %>%
knitr::kable(digits=2)
rowname auc acc sens spec ppv npv fpr
cnn_4 0.66 0.53 0.79 0.42 0.38 0.81 0.58 cnn_10 0.65 0.54 0.79 0.44 0.39 0.82 0.56 cnn_8 0.56 0.41 0.79 0.23 0.32 0.71 0.77
Save Classifier Lists As Package Data
classifiers <- list(
"classifier_points" = cPs,
"classifier_curves" = cCs
)
str(classifiers)
#> List of 2
#> $ classifier_points:List of 7
#> ..$ nlp_4 :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] TRUE TRUE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ nlp_8 :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE FALSE FALSE FALSE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ nlp_10 :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE FALSE FALSE FALSE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ javin :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ eko :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] TRUE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ mike :Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] TRUE FALSE FALSE TRUE TRUE TRUE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ priority:Formal class 'ClassifierPoint' [package "AnalysisToolkit"]..
#> .. .. ..@ Y_: logi [1:180] FALSE FALSE FALSE FALSE FALSE FALSE ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> $ classifier_curves:List of 3
#> ..$ cnn_8 :Formal class 'ClassifierCurve' [package "AnalysisToolkit"] w..
#> .. .. ..@ pY: num [1:180] 0.25347 0.65845 0.40168 0.76503 0.00696 ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ cnn_10:Formal class 'ClassifierCurve' [package "AnalysisToolkit"] w..
#> .. .. ..@ pY: num [1:180] 0.196 0.705 0.113 0.117 0.106 ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
#> ..$ cnn_4 :Formal class 'ClassifierCurve' [package "AnalysisToolkit"] w..
#> .. .. ..@ pY: num [1:180] 0.992 1 1 1 1 ...
#> .. .. ..@ Y : logi [1:180] FALSE FALSE TRUE TRUE TRUE TRUE ...
#> .. .. ..@ id: chr [1:180] "7862738" "7863668" "7872384" "7873551" ...
# devtools::use_data(classifiers)
Sys Info
Sys.info()
#> sysname release version nodename
#> "Windows" "10 x64" "build 16299" "MARCUSDESKTOP"
#> machine login user effective_user
#> "x86-64" "Marcus User" "Marcus User" "Marcus User"
search()
#> [1] ".GlobalEnv" "package:bindrcpp"
#> [3] "package:pander" "package:knitr"
#> [5] "package:vizR" "package:cowplot"
#> [7] "package:AnalysisToolkit" "package:broom"
#> [9] "package:ProjUtils" "package:MyUtils"
#> [11] "package:forcats" "package:purrr"
#> [13] "package:shiny" "package:magrittr"
#> [15] "package:rebus" "package:glue"
#> [17] "package:stringr" "package:tidyr"
#> [19] "package:dplyr" "package:plyr"
#> [21] "package:readr" "package:beepr"
#> [23] "package:ggplot2" "package:stats"
#> [25] "package:graphics" "package:grDevices"
#> [27] "package:utils" "package:datasets"
#> [29] "package:testthat" "package:assertive"
#> [31] "package:assertthat" "package:methods"
#> [33] "Autoloads" "package:base"
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.