in prep for cran submission, fix typos, reduce example run time, fix broken link and replace rawgit links with https://cdn.staticaly.com/gh/ because rawgit will be down very soon.

Yue-Jiang · Yue-Jiang · commit 09450c44d504 · 2019-09-17T14:12:19.000-07:00
diff --git a/.Rbuildignore b/.Rbuildignore
@@ -5,3 +5,5 @@ materials*
 ^_pkgdown\.yml$
 ^\.travis\.yml$
 ^cran-comments\.md$
+^doc$
+^Meta$
diff --git a/.gitignore b/.gitignore
@@ -1,3 +1,5 @@
 .Rproj.user
 .Rhistory
 .RData
+doc
+Meta
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: randomForestExplainer
 Title: Explaining and Visualizing Random Forests in Terms of Variable Importance
-Version: 0.9
+Version: 0.10.0
 Authors@R: c(
   person("Aleksandra", "Paluszynska", email = "ola.paluszynska@gmail.com", role = c("aut")),
   person("Przemyslaw", "Biecek", email = "przemyslaw.biecek@gmail.com", role = c("aut","ths")),
@@ -18,13 +18,13 @@ Imports:
   GGally (>= 1.3.0),
   ggplot2 (>= 2.2.1),
   ggrepel (>= 0.6.5),
-  MASS (>= 7.3.47),
   randomForest (>= 4.6.12),
   ranger(>= 0.9.0),
   reshape2 (>= 1.4.2),
   rmarkdown (>= 1.5)
 Suggests:
     knitr,
+    MASS (>= 7.3.47),
     testthat
 VignetteBuilder: knitr
 RoxygenNote: 6.1.1
diff --git a/NEWS.md b/NEWS.md
@@ -0,0 +1,8 @@
+# randomForestExplainer 0.10.0
+## New features
+* Added support for ranger forests.
+* Added support for unsupervised randomForest.
+* Added tests for most functions.
+
+## Bug fixes
+* More intuitive error message for explain_forest when local importance is absent.
diff --git a/R/measure_importance.R b/R/measure_importance.R
@@ -246,7 +246,7 @@ measure_importance.ranger <- function(forest, mean_sample = "top_trees", measure
 #' @param importance_frame A result of using the function measure_importance() to a random forest or a randomForest object
 #' @param measures A character vector specifying the measures of importance to be used
 #' @param k The number of variables to extract
-#' @param ties_action One of three: c("none", "all", "draw"); specifies which variables to pick when ties occur. When set to "none" we may get less than k variables, when "all" whe may get more and "draw" makes us get exactly k.
+#' @param ties_action One of three: c("none", "all", "draw"); specifies which variables to pick when ties occur. When set to "none" we may get less than k variables, when "all" we may get more and "draw" makes us get exactly k.
 #'
 #' @return A character vector with names of k variables with highest sum of rankings
 #'
diff --git a/R/min_depth_distribution.R b/R/min_depth_distribution.R
@@ -39,8 +39,8 @@ calculate_tree_depth_ranger <- function(frame){
 #' @return A data frame with the value of minimal depth for every variable in every tree
 #'
 #' @examples
-#' min_depth_distribution(randomForest::randomForest(Species ~ ., data = iris))
-#' min_depth_distribution(ranger::ranger(Species ~ ., data = iris))
+#' min_depth_distribution(randomForest::randomForest(Species ~ ., data = iris, ntree = 100))
+#' min_depth_distribution(ranger::ranger(Species ~ ., data = iris, num.trees = 100))
 #'
 #' @export
 min_depth_distribution <- function(forest){
diff --git a/R/min_depth_interactions.R b/R/min_depth_interactions.R
@@ -110,11 +110,11 @@ min_depth_interactions_values_ranger <- function(forest, vars){
 #' Calculate mean conditional minimal depth with respect to a vector of variables
 #'
 #' @param forest A randomForest object
-#' @param vars A character vector with variables with respect to which conditional minimal depth will be calculated; by defalt it is extracted by the important_variables function but this may be time consuming
+#' @param vars A character vector with variables with respect to which conditional minimal depth will be calculated; by default it is extracted by the important_variables function but this may be time consuming
 #' @param mean_sample The sample of trees on which conditional mean minimal depth is calculated, possible values are "all_trees", "top_trees", "relevant_trees"
 #' @param uncond_mean_sample The sample of trees on which unconditional mean minimal depth is calculated, possible values are "all_trees", "top_trees", "relevant_trees"
 #'
-#' @return A data frame with each observarion giving the means of conditional minimal depth and the size of sample for a given interaction
+#' @return A data frame with each observation giving the means of conditional minimal depth and the size of sample for a given interaction
 #'
 #' @examples
 #' forest <- randomForest::randomForest(Species ~ ., data = iris, ntree = 100)
diff --git a/README.md b/README.md
@@ -5,9 +5,9 @@
 [![codecov](https://codecov.io/gh/ModelOriented/randomForestExplainer/branch/master/graph/badge.svg)](https://codecov.io/gh/ModelOriented/randomForestExplainer)
 [![DOI](https://zenodo.org/badge/97007621.svg)](https://zenodo.org/badge/latestdoi/97007621)
 
-A set of tools to understand what is happening inside a Random Forest. A detailed discussion of the package and importance measures it implements can be found here: [Master thesis on randomForestExplainer](https://rawgit.com/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer_Master_thesis.pdf).
+A set of tools to understand what is happening inside a Random Forest. A detailed discussion of the package and importance measures it implements can be found here: [Master thesis on randomForestExplainer](https://cdn.staticaly.com/gh/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer_Master_thesis.pdf).
 
-## Instalation
+## Installation
 
 randomForestExplainer can be installed from [CRAN](https://cran.r-project.org/package=randomForestExplainer) as follows:
 
@@ -26,12 +26,12 @@ library(randomForestExplainer)
 
 ## Vignette
 
-* [Understanding random forests with randomForestExplainer](https://rawgit.com/ModelOriented/randomForestExplainer/master/inst/doc/randomForestExplainer.html)
+* [Understanding random forests with randomForestExplainer](https://cdn.staticaly.com/gh/ModelOriented/randomForestExplainer/master/inst/doc/randomForestExplainer.html)
 
 ## Cheatsheets
 
 * [A one-page summary](https://github.com/ModelOriented/randomForestExplainer/blob/master/materials/cheatsheet.pdf)
 
 ## Examples
 
-* [Initial vignette for glioblastoma data](https://rawgit.com/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer/inst/doc/randomForestExplainer.html)
+* [Initial vignette for glioblastoma data](https://cdn.staticaly.com/gh/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer/inst/doc/randomForestExplainer.html)
diff --git a/man/important_variables.Rd b/man/important_variables.Rd
diff --git a/man/min_depth_distribution.Rd b/man/min_depth_distribution.Rd
diff --git a/man/min_depth_interactions.Rd b/man/min_depth_interactions.Rd
diff --git a/vignettes/randomForestExplainer.Rmd b/vignettes/randomForestExplainer.Rmd
@@ -16,7 +16,7 @@ knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE, fig.width =
 
 # Introduction
 
-This vignette demonstrates how to use the randomForestExplainer package. We will use the `Boston` from package `MASS`. In fact, the development of randomForestExplainer was motivated by problems that include lots of predictors and not many observations. However, as they usually require growing large forests and are computationally intensive, we use the small `Boston` data set here and encourage you to follow our analysis of such large data set concerning glioblastoma here: [initial vignette](https://rawgit.com/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer/inst/doc/randomForestExplainer.html).
+This vignette demonstrates how to use the randomForestExplainer package. We will use the `Boston` from package `MASS`. In fact, the development of randomForestExplainer was motivated by problems that include lots of predictors and not many observations. However, as they usually require growing large forests and are computationally intensive, we use the small `Boston` data set here and encourage you to follow our analysis of such large data set concerning glioblastoma here: [initial vignette](https://cdn.staticaly.com/gh/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer/inst/doc/randomForestExplainer.html).
 
 We will use the following packages:
 
@@ -79,7 +79,7 @@ The function `plot_min_depth_distribution` offers three possibilities when it co
 
 Note that the $x$-axis ranges from zero trees to the maximum number of trees in which any variable was used for splitting ($\tilde{B}$) which is in this case equal to 500 and is reached by all variables plotted.
 
-The ordering of variables in our plot by their mean minimal depth seems quite accurate when we look at the distribution of minimal depth, though one could argue that for example `indus` should be ranked higherthan `dis` as the latter is never used for splitting at the root. Usually we would obtain different orderings when changing the `mean_sample` option but this is not the case if variables are used for splitting in all trees as this option only influences how and whether missing values are treated. The default option, `"top_trees"`, penalizes missing values and this penalization makes the interpretation of the values less obvious -- to address that we can calculate the mean minimal depth only using non-missing observations (`mean_sample = "relevant_trees"`). For forests with many variables with a lot of missing observations we should always consider adding the `min_no_of_trees` option so that only variables used for splitting in at least the declared number of trees will be considered for the plot. This allows us to avoid selecting variables that have been by chance used for splitting e.g., only once but at the root (their mean would be equal to 0). However, in our case we can simply increase the `k` parameter to plot all trees:
+The ordering of variables in our plot by their mean minimal depth seems quite accurate when we look at the distribution of minimal depth, though one could argue that for example `indus` should be ranked higher than `dis` as the latter is never used for splitting at the root. Usually we would obtain different orderings when changing the `mean_sample` option but this is not the case if variables are used for splitting in all trees as this option only influences how and whether missing values are treated. The default option, `"top_trees"`, penalizes missing values and this penalization makes the interpretation of the values less obvious -- to address that we can calculate the mean minimal depth only using non-missing observations (`mean_sample = "relevant_trees"`). For forests with many variables with a lot of missing observations we should always consider adding the `min_no_of_trees` option so that only variables used for splitting in at least the declared number of trees will be considered for the plot. This allows us to avoid selecting variables that have been by chance used for splitting e.g., only once but at the root (their mean would be equal to 0). However, in our case we can simply increase the `k` parameter to plot all trees:
 
 ```{r fig.height = 7}
 plot_min_depth_distribution(min_depth_frame, mean_sample = "relevant_trees", k = 15)
@@ -139,7 +139,7 @@ Observe the marked negative relation between `times_a_root` and `mean_min_depth`
 plot_multi_way_importance(importance_frame, x_measure = "mse_increase", y_measure = "node_purity_increase", size_measure = "p_value", no_of_labels = 5)
 ```
 
-As in the previous plot, the two measures used as coordinates seem correlated, but in this case this is somewhat more surprising as one is connected to the structure of the forest and the other to its prediction, whereas in the previous plot both measures reflected the structure. Also, in this plot we see that although `lstat` and `rm` are similar in terms of node purity increase and $p$-value, the former is markedly better if we look at the increase in MSE. Interestingly, `nox` and `indus` are guite good when it comes to the two measures reflected on the axes, but are not significant according to our $p$-value, which is a derivative of the number of nodes that use a variable for splitting.
+As in the previous plot, the two measures used as coordinates seem correlated, but in this case this is somewhat more surprising as one is connected to the structure of the forest and the other to its prediction, whereas in the previous plot both measures reflected the structure. Also, in this plot we see that although `lstat` and `rm` are similar in terms of node purity increase and $p$-value, the former is markedly better if we look at the increase in MSE. Interestingly, `nox` and `indus` are quite good when it comes to the two measures reflected on the axes, but are not significant according to our $p$-value, which is a derivative of the number of nodes that use a variable for splitting.
 
 ## Compare measures using ggpairs
 
@@ -223,7 +223,7 @@ The `explain_forest()` function is the flagship function of the `randomForestExp
 explain_forest(forest, interactions = TRUE, data = Boston)
 ```
 
-To see the resulting HTML document click here: [Boston forest summary](https://rawgit.com/geneticsMiNIng/BlackBoxOpener/master/examples/Boston_forest_explained.html)
+To see the resulting HTML document click here: [Boston forest summary](https://cdn.staticaly.com/gh/geneticsMiNIng/BlackBoxOpener/master/examples/Boston_forest_explained.html)
 
-For additional examples see: [initial vignette](https://rawgit.com/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer/inst/doc/randomForestExplainer.html).
+For additional examples see: [initial vignette](https://cdn.staticaly.com/gh/geneticsMiNIng/BlackBoxOpener/master/randomForestExplainer/inst/doc/randomForestExplainer.html).
 

-Original file line number
+Diff line change
@@ @@ -1,3 +1,5 @@ @@
 .Rproj.user
 .Rhistory
 .RData
 +doc
 +Meta