Abstract #

We report a new release of the self-explainable AutoML software AMLBID. The software package is a meta-learning based AutoML decision support system that assists machine learning (ML) experts and neophytes in building well performing ML pipelines and makes the outcomes transparent and interpretable. The release 2.0 of the software introduces an enhanced efficiency of algorithms recommendation. The performance improvement is mainly achieved by the integration of the AeKNN meta-model. This implementation makes datasets meta-features more informative and further improve the accuracy. The release 2.0 also introduces the standalone AMLBIDesc, a desktop version of the AMLBID software which makes the tool more accessible to non-expert users.

Citation #

@article{Garouani2023_softx,
  doi = {10.1016/j.softx.2023.101444},
  title = {Version [2.0]- [AMLBID: An auto-explained Automated Machine Learning tool for Big Industrial Data]},
  journal = {SoftwareX},
  volume = {23},
  pages = {101444},
  year = {2023},
  issn = {2352-7110},
  author = {Moncef Garouani and Mourad Bouneffa and Adeel Ahmad and Mohamed Hamlich},
}