Datasets & Source Code

Datasets

Source Code

  • Label Ranking Ensembles

    The Label Ranking Ensembles source code corresponds to the experimental flow of section 4 in our paper: “Beyond Majority: Label Ranking Ensembles based on Voting Rules”.

    The source code can be downloaded from:
    http://bigdatalab.tau.ac.il/shared_resources/source_code/label_ranking_source_code.zip

    The source code includes two parts:

    1. Java code for partitioning the datasets, sampling the data using bagging and training weak learners (LRT or RPC).

    2. Python code for prediction using various voting rules as aggregation methods for the ensembles, and evaluation using the kendall-tau-b measure.

    Upon use, please make sure to cite:

    Havi Werbin-Ofir, Lihi Dery and Erez Shmueli. 2019.  “Beyond Majority: Label Ranking Ensembles based on Voting Rules”. Expert Systems with Applications. 136: 50-61.

Reports