Local set computation for multi-label data sets

Instance selection algorithms for MEKA. The algorithms LSSm and LSBo have been adapted to multi-label learning.

The original MEKA software is necessary: https://github.com/Waikato/meka

• HD-LSSm: Local set-based smoother for multi-label.
• HD-LSBo: Local set border selector for multi-label.

Citation policy

Á. Arnaiz-González, J-F. Díez Pastor, Juan J. Rodríguez, C. García Osorio. Local sets for multi-label instance selection. Applied Soft Computing, 68, 651-666. doi: 10.1016/j.asoc.2018.04.016

@article{ArnaizGonzalez2018,
  title = "Local sets for multi-label instance selection",
  journal = "Applied Soft Computing ",
  volume = "68",
  pages = "651 - 666",
  year = "2018",
  issn = "1568-4946",
  doi = "10.1016/j.asoc.2018.04.016",
  author = "\'{A}lvar Arnaiz-Gonz\'{a}lez and Jos\'{e} F. D\'{i}ez-Pastor and Juan J. Rodr\'{i}guez and C\'{e}sar Garc\'{i}a-Osorio"
}

The paper is available online and accessible for free until July 11, 2018 through this link https://authors.elsevier.com/a/1X5Q15aecSZbPu

Contributions

The algorithms are based on LSSm and LSBo: Leyva, E., González, A., & Pérez, R. (2015). Three new instance selection methods based on local sets: A comparative study with several approaches from a bi-objective perspective. Pattern Recognition, 48(4), 1523-1537.