NEWS.md

In MetricsWeighted: Weighted Metrics and Performance Measures for Machine Learning

MetricsWeighted 1.0.3

Improvements

• logLoss() and deviance_poisson() are now less picky about predictions of 0 (or 1).

MetricsWeighted 1.0.2

Documentation

• Fix Latex problem in MacOS help files.
• Slight improvements in docu.

MetricsWeighted 1.0.1

Documentation

• Much more compact help files
• Using Latex in help files

MetricsWeighted 1.0.0

This release bumps the package to stable version 1.0.0.

Maintenance

• Reorganization of code
• Switch to package::function() notation

MetricsWeighted 0.5.5

Maintenance

• Use github workflows
• Add github pages
• Better README

Dependencies

• Removed {dplyr} from suggested packages.

MetricsWeighted 0.5.4

This is a maintainance update without any code change.

• Fixed problematic unit tests.
• Internal restructuring how package content is being generated.

MetricsWeighted 0.5.3

• Added unit tests
• Added option reference_mean to r_squared() functions. This allows clean out-of-sample applications.
• Added Murphy diagrams

MetricsWeighted 0.5.2

Maintainance release. rmarkdown is now an explicit "Suggested" package.

MetricsWeighted 0.5.1

New function

• weighted_cor to calculate weighted correlation between actual and observed values.

• prop_within to calculate weighted proportion of predictions within a tolerance around actual values.

Requirements

• Reduced minimal R version from 3.5 to 3.1.

Documentation

• Clarified that R-squared is calculated with respect to the null model calculated from the actual values.

MetricsWeighted 0.5.0

New functions

• Elementary scoring functions for expectiles and quantiles.

• multi_metric: A way to create a named list of performance measures parametrized by a parameter.

MetricsWeighted 0.4.0

New functions

Added the following convenience wrappers to r_squared.

• r_squared_poisson

• r_squared_gamma

• r_squared_bernoulli

MetricsWeighted 0.3.0

New function

Added function weighted_var to calculate variance weighted by sampling weights.

MetricsWeighted 0.2.0

• Improvement of documentation and examples.

• Better handling of Tweedie special cases.

• More strict error handling.

• Added median absolute error (and weighted_median, weighted_quantile)

MetricsWeighted 0.1.0

Initial release.

MetricsWeighted documentation built on Nov. 16, 2023, 5:09 p.m.