Laurae (version 0.0.0.9001)

partial_dep.obs_all: Partial Dependency Observation, Contour (multiple observations)

Description

This function computes partial dependency of a supervised machine learning model over a range of values for multiple observation. Does not work for multiclass problems! Beware as the number of data might explode, especially if you have many different data. The amount of observations you will end up explodes quickly and is linear to the sum of each (max(unique value of feature, accuracy)). For instance, 100 features with 50 unique values and 1000 observations ends up as 5,000,000 observations!!!!! (5000 times the initial dataset)

Usage

partial_dep.obs_all(model, predictor, data, observation,
  column = colnames(data), accuracy = min(length(data), 10),
  exact_only = TRUE, label_name = "Target", comparator_name = "Evolution")

Arguments

model

Type: unknown. The model to pass to predictor.

predictor

Type: function(model, data). The predictor function which takes a model and data as inputs, and return predictions. data is provided as data.table for maximum performance.

data

Type: data.table (mandatory). The data we need to use to sample from for the partial dependency.

observation

Type: data.table (mandatory). The observation we want to get partial dependence from. You can put the same as data if you wish to get partial dependence from it.

column

Type: character. The column we want partial dependence from. You can specify two or more column as a vector, but it is highly not recommended to go for a lot of columns because the complexity is linearly multiplicative, think as O*length(column)*accuracy. For instance, accuracy = 100, length(column) = 100, and nrow(data) = 1000000 leads to 1e+10 theoretical observations, which could explode the memory of any computer.

accuracy

Type: integer. The accuracy of the partial dependence from, exprimed as number of sampled points by percentile of the column from the the data. Defaults to min(length(data), 10), which means either 10 samples or all samples of data if the latter has less than 10 observations.

exact_only

Type: logical. Whether to select only exact values for data sampling. Defaults to TRUE.

label_name

Type: character. The column name given to the predicted values in the output table. Defaults to "Target", this assumes you do not have a column called "Target" in your column vector.

comparator_name

Type: character. The column name given to the evolution value ("Increase", "Fixed", "Decrease") in the output table. Defaults to "Evolution", this assumes you do not have a column called "Evolution" in your column vector.

Value

A list with different elements: grid_init for the grid before expansion, grid_exp for the expanded grid with predictions, preds for the predictions, and obs for the original predictions on data

Examples

## Not run: ------------------------------------
# # Let's load a dummy dataset
# data(mtcars)
# setDT(mtcars) # Transform to data.table for easier manipulation
# 
# # We train a xgboost model on 31 observations, keep last to analyze later
# set.seed(0)
# xgboost_model <- xgboost(data = data.matrix(mtcars[, -1]),
#                          label = mtcars$mpg,
#                          nrounds = 20)
# 
# # Perform partial dependence grid prediction to analyze the behavior of the 32th observation
# # We want to check how it behaves with:
# # => horsepower (hp)
# # => number of cylinders (cyl)
# # => transmission (am)
# # => number of carburetors (carb)
# preds_partial <- partial_dep.obs_all(model = xgboost_model,
#                                      predictor = predictor_xgb, # Default for xgboost
#                                      data = mtcars[, -1], # train data
#                                      observation = mtcars[, -1], # train data
#                                      # when column is not specified => all columns
#                                      accuracy = 20, # Up to 20 unique values per column
#                                      exact_only = TRUE, # Not allowing approximations,
#                                      label_name = "mpg", # Label is supposed "mpg"
#                                      comparator_name = "evo") # Comparator +/-/eq for analysis
# 
# # How many observations? 3360, that's a lot coming from original 32 observations.
# nrow(preds_partial$grid_exp)
# 
# # How many observations analyzed per column?
# summary(preds_partial$grid_init)
# #      Length Class  Mode   
# # cyl   3     -none- numeric
# # disp 19     -none- numeric
# # hp   16     -none- numeric
# # drat 16     -none- numeric
# # wt   19     -none- numeric
# # qsec 19     -none- numeric
# # vs    2     -none- numeric
# # am    2     -none- numeric
# # gear  3     -none- numeric
# # carb  6     -none- numeric
# 
# # Great plotting skills!
# partial_dep.plot(preds_partial$grid_exp,
#                  backend = c("plotly", "line"),
#                  label_name = "mpg",
#                  comparator_name = "evo")
# 
# # Get statistics to analyze fast
# partial_dep.feature(preds_partial$grid_exp, metric = "emp", in_depth = FALSE)
# 
# # Get statistics to analyze, but is very slow when there is large data
# # Note: unreliable for large amount of observations due to asymptotic infinites
# partial_dep.feature(preds_partial$grid_exp, metric = "emp", in_depth = TRUE)
## ---------------------------------------------