docs & pkgdown

NEWS.md

@@ -1,6 +1,7 @@
 # live 1.5.9
-* Dropped old interface.
 
+* Dropped old interface.
+* Improved distance calculations.
 
 # live 1.5.8
 

R/app.R

@@ -22,7 +22,8 @@ live_shiny <- function(train_data, black_box_model, target, explained_data = tra
              sliderInput("size", "Size of simulated dataset",
                          min = 100, max = 10000, step = 100,
                          round = T, value = 1000),
-             selectInput("method", "Sampling method", choices = c("live", "lime"),
+             selectInput("method", "Sampling method",
+                         choices = c("live", "permute", "normal"),
                          selected = "live"),
              checkboxInput("standardize", "Center predictors"),
              checkboxInput("selection", "Variable selection"),

R/explore.R

@@ -19,7 +19,6 @@
 #' @return list of class "live_explorer" consisting of
 #' \item{data}{Dataset generated by sample_locally function with response variable.}
 #' \item{target}{Name of the response variable.}
-#' \item{model}{Black box model which is being explained.}
 #' \item{explained_instance}{Instance that is being explained.}
 #' \item{sampling_method}{Name of used sampling method}
 #' \item{fixed_variables}{Names of variables which were not sampled}
@@ -32,8 +31,7 @@
 #' dataset_for_local_exploration <- sample_locally(data = wine,
 #'                                                 explained_instance = wine[5, ],
 #'                                                 explained_var = "quality",
-#'                                                 size = 50,
-#'                                                 standardise = TRUE)
+#'                                                 size = 50)
 #' }
 #'
 
@@ -121,11 +119,13 @@ give_predictions <- function(data, black_box, explained_var, similar, predict_fu
 #'
 #' @examples
 #' \dontrun{
-#' local_exploration1 <- add_predictions(wine, dataset_for_local_exploration,
-#'                                       black_box_model = "regr.svm")
+#' # Train a model inside add_predictions call.
+#' local_exploration1 <- add_predictions(dataset_for_local_exploration,
+#'                                       black_box_model = "regr.svm",
+#'                                       data = wine)
 #' # Pass trained model to the function.
 #' svm_model <- svm(quality ~., data = wine)
-#' local_exploration2 <- add_predictions(wine, dataset_for_local_exploration,
+#' local_exploration2 <- add_predictions(dataset_for_local_exploration,
 #'                                       black_box_model = svm_model)
 #' }
 #'

R/live.R

@@ -2,7 +2,7 @@
 #'
 #' This package aims to help locally fit and visualize interpretable models similarly to LIME methodology.
 #' Interface provided by mlr package is used. Tools are provided to create a simulated dataset of
-#' similar observations, fit a chosen white box models (GLM and CART in particular) and visualize
+#' similar observations, fit chosen white box models (GLM and CART in particular) and visualize
 #' them. The methodology is based on Tulio Ribeiro, Singh, Guestrin (2016) <doi:10.1145/2939672.2939778>.
 #' More details can be found in Staniak, Biecek (2018) https://arxiv.org/abs/1804.01955.
 #'

R/plot.R

@@ -1,15 +1,5 @@
-#' Waterfall plot or forestplot for lm/glm explanations.
-#'
-#' @param plot_type Chr, "forest" or "waterfall" depending
-#'                  on which type of plot is to be created.
-#' @param fitted_model glm or lm object.
-#' @param explained_instance Observation around which model was fitted.
-#' @param classif logical, if TRUE, probabilities will be plotted 
-#'
 #' @importFrom graphics plot
-#'
-#' @return plot (ggplot2 or lattice)
-#'
+
 
 plot_regression <- function(plot_type, fitted_model, explained_instance, classif) {
   if(plot_type == "forest") {