| Title: | Computes Conformal p-Values |
|---|---|
| Description: | Computes marginal conformal p-values using conformal prediction in binary classification tasks. Conformal prediction is a framework that augments machine learning algorithms with a measure of uncertainty, in the form of prediction regions that attain a user-specified level of confidence. This package specifically focuses on providing conformal p-values that can be used to assess the confidence of the classification predictions. For more details, see Tyagi and Guo (2023) <https://proceedings.mlr.press/v204/tyagi23a.html>. |
| Authors: | Chhavi Tyagi [aut, cre] |
| Maintainer: | Chhavi Tyagi <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 0.1.0 |
| Built: | 2025-03-08 03:30:12 UTC |
| Source: | https://github.com/cran/conformalpvalue |
This function calculates conformal p-values based of binary class labels for test data.
conformal_pvalues(train_data, calib_data, test_data, target_col, method)conformal_pvalues(train_data, calib_data, test_data, target_col, method)
train_data |
A data frame containing the training data with the target variable. |
calib_data |
A data frame containing the calibration data with the target variable. |
test_data |
A data frame containing the test data. |
target_col |
The name of the target variable column. |
method |
A character string specifying the classification method to use. Options are 'naiveBayes', 'svm', and 'glm'. This function trains a Naive Bayes classifier, computes non-conformity scores on the calibration data and test data, and calculates conformal p-values of both classes "0" and "1" using the conformal prediction for a binary classification problem. |
A matrix containing p-values for each test case and class.
# Create dummy train_data, calib_data, and test_data train_data <- data.frame( x1 = as.numeric(rnorm(50, 1, 2)), x2 = as.numeric(rnorm(50, 2.5, 3)), target = as.factor(rbinom(50, 1, 0.5)) ) calib_data <- data.frame( x1 = as.numeric(rnorm(50, 1, 2)), x2 = as.numeric(rnorm(50, 2.5, 3)), target = as.factor(rbinom(50, 1, 0.5)) ) test_data <- data.frame( x1 = as.numeric(rnorm(50, 1, 2)), x2 = as.numeric(rnorm(50, 2.5, 3)) ) p_values <- conformal_pvalues(train_data, calib_data, test_data, target="target", method="svm")# Create dummy train_data, calib_data, and test_data train_data <- data.frame( x1 = as.numeric(rnorm(50, 1, 2)), x2 = as.numeric(rnorm(50, 2.5, 3)), target = as.factor(rbinom(50, 1, 0.5)) ) calib_data <- data.frame( x1 = as.numeric(rnorm(50, 1, 2)), x2 = as.numeric(rnorm(50, 2.5, 3)), target = as.factor(rbinom(50, 1, 0.5)) ) test_data <- data.frame( x1 = as.numeric(rnorm(50, 1, 2)), x2 = as.numeric(rnorm(50, 2.5, 3)) ) p_values <- conformal_pvalues(train_data, calib_data, test_data, target="target", method="svm")