TaskDesc: Description object for task.
Description
Description object for task, encapsulates basic properties of the task
without having to store the complete data set.
Details
Object members:
- id [
character(1)] - Id string of task.
- type [
character(1)] - Type of task, “classif” for classification,
“regr” for regression, “surv” for survival and “cluster” for
cluster analysis, “costsens” for cost-sensitive classification, and
“multilabel” for multilabel classification.
- target [
character(0) | character(1) | character(2) | character(n.classes)] -
Name(s) of the target variable(s).
For “surv” these are the names of the survival time and event columns, so it has length 2.
For “costsens” it has length 0, as there is no target column, but a cost matrix instead.
For “multilabel” these are the names of logical columns that indicate whether a
class label is present and the number of target variables corresponds to the number of
classes.
- size [
integer(1)] - Number of cases in data set.
- n.feat [
integer(2)] - Number of features, named vector with entries:
“numerics”, “factors”, “ordered”.
- has.missings [
logical(1)] - Are missing values present?
- has.weights [
logical(1)] - Are weights specified for each observation?
- has.blocking [
logical(1)] - Is a blocking factor for cases available in the task?
- class.levels [
character] - All possible classes.
Only present for “classif”, “costsens”, and “multilabel”.
- positive [
character(1)] - Positive class label for binary classification.
Only present for “classif”, NA for multiclass.
- negative [
character(1)] - Negative class label for binary classification.
Only present for “classif”, NA for multiclass.
- censoring [
character(1)] - Censoring type for survival analysis.
Only present for “surv”, one of “rcens” for right censored data,
“lcens” for left censored data, and “icens” for interval censored
data.