pp_score: pp_score

In Gulde91/Predictive_power_score: What the Package Does (One Line, Title Case)

Description

Calculates predictive power score for x predicts y. Where x can be multiple features in df. This is a light (wip) implementation which for now only supports binary classification and x as numeric features.

Usage

pp_score(
  df,
  x,
  y,
  metric = "roc_auc",
  sample_size = NULL,
  cv_folds = 5,
  repeated_cv = 1
)

Arguments

df	data.frame
x	str name of features in df
y	str target feature
metric	str Machine learning metric to evaluate on. Choose between roc_auc, pr_auc, F1
sample_size	int random sample taken from df to speed up calculations. If NULL all samples are used.
cv_folds	number of cross validations folds
repeated_cv	int number of repeated cross validations

Value

A list with the predictive power score for each feature in x

Gulde91/Predictive_power_score documentation built on Dec. 31, 2020, 12:06 p.m.