mmetrics

Overview

mmetrics is a grammar of metrics calculation, providing a consistent set of functions that help you quantitatively assess the performance of statistical models:

• confusion_scores and multiclass_confusion_scores provides confusion matrices given model predictions and ground truth labels.

• accuracy function (binary_acc and multiclass_acc) estimates the correctness of predicted classification labels.

• precision function (binary_precision and multiclass_precision) measures the accuracy of positive predictions.

• recall function (binary_recall and multiclass_recall) measures the completeness of positive predictions.

• f1 score computation (binary_f1 and multiclass_f1) gives a balanced view of precision and recall scores, reflecting how well a classification model is to assign each observation to their corresponding calss.

Installation

install.packages("mmetrics")

Usage

Please refer to doc/Tutorial.html for example usage or consult function help page.