MultiLevelIPF: Implementation of algorithms that extend IPF to nested structures

The IPF algorithm operates on count data. This package offers implementations for several algorithms that extend this to nested structures: "parent" and "child" items for both of which constraints can be provided.

Details

To use this package, you need to:

1. Specify your fitting problem with fitting_problem()

2. Optionally, convert the fitting problem to a structure that can be processed by the algorithms with flatten_ml_fit_problem(); this is helpful if you want to run the same fitting problem with multiple algorithms and compare results.

3. Compute weights with one of the algorithms provided in this package with ml_fit() or one of the specialized functions

4. Analyze weights or residuals, e.g. with compute_margins()

See also

Useful links:

• https://mlfit.github.io/MultiLevelIPF

• https://github.com/mlfit/MultiLevelIPF

• Report bugs at https://github.com/mlfit/MultiLevelIPF/issues

Author

Maintainer: Kirill Müller mlfit+r@mailbox.org

Other contributors:

• Amarin Siripanich amarin.siri@gmail.com [contributor]