calculate_overall_feature_importance: Calculating feature importances across trajectories

In dynverse/dynfeature: Feature Importance for Dynamic Processes

Calculating feature importances across trajectories

Description

Uses the feature importance measures of ranger or caret. calculate_overall_feature_importance calculates the importance for the whole trajectory, calculate_milestone_feature_importance calculates it for individual milestones (e.g. branching points)

Usage

calculate_branch_feature_importance(
  trajectory,
  expression_source = "expression",
  fi_method = fi_ranger_rf_lite(),
  verbose = FALSE
)

calculate_branching_point_feature_importance(
  trajectory,
  expression_source = "expression",
  milestones_oi = trajectory$milestone_ids,
  fi_method = fi_ranger_rf_lite(),
  verbose = FALSE
)

calculate_cell_feature_importance(
  trajectory,
  expression_source = "expression",
  fi_method = fi_ranger_rf_lite(),
  verbose = FALSE
)

calculate_milestone_feature_importance(
  trajectory,
  expression_source = "expression",
  milestones_oi = NULL,
  fi_method = fi_ranger_rf_lite(),
  verbose = FALSE
)

calculate_overall_feature_importance(
  trajectory,
  expression_source = "expression",
  fi_method = fi_ranger_rf_lite(),
  verbose = FALSE
)

calculate_waypoint_feature_importance(
  trajectory,
  expression_source = "expression",
  waypoints = NULL,
  fi_method = fi_ranger_rf_lite(),
  verbose = FALSE
)

Arguments

trajectory	A trajectory object containing expression values and a trajectory.
expression_source	The expression data matrix, with features as columns. If a matrix is provided, it is used as is. If a character is provided, trajectory[[expression_source]] should contain the matrix. If a function is provided, that function will be called in order to obtain the expression (useful for lazy loading).
fi_method	A feature importance method. Default: fi_ranger_rf_lite(). Check ?fi_methods for a full list of available feature importance methods.
verbose	Whether to print out extra information.
milestones_oi	The milestone(s) for which to calculate feature importance
waypoints	The waypoints, optional

Value

A data frame with two or more columns, feature_id, and importance. feature_id is a column in the trajectory expression matrix. Additional columns may be available depending on the function called.

Examples

library(dynwrap)
data(example_trajectory)

calculate_overall_feature_importance(example_trajectory)

dynverse/dynfeature documentation built on March 3, 2023, 1:44 p.m.