Unveiling the Power of Disturbing Neighbors: A Comparative Study of Ensemble Methods for Semi-Supervised Learning

This is the official repository of the paper "Unveiling the Power of Disturbing Neighbors: A Comparative Study of Ensemble Methods for Semi-Supervised Learning".

Citation

If you use the code, please cite the following paper:

PENDING PUBLICATION

Structure of the repository

This repository contains:

• The source code of the Disturbing Neighbors (DN) algorithm -> From the repository jlgarridol/admirable-methods (License: BSD-3-Clause)
• The source code of the experiments
• The datasets used in the experiments
• The results of the experiments
• The Jupyter Notebooks used to generate the tests and the plots of the paper

The repository is structured as follows:

• datasets/: Contains the datasets used in the experiments compressed in a tar.xz file. They are from UCI Machine Learning Repository and can be downloaded from here. The format of the datasets are .csv.
• results/: Contains the results of the experiments and the Jupyter Notebooks used to generate the tests and the plots of the paper. The resutls are two files: dn_results.pkl and dn_f1_results.pkl. The first file contains the accuracy of the experiments and the second file contains the F1-score of the experiments. Both are in dill format.
• experiments/: Contains the source code of the experiments. The experiments are launched over the framework.py file and configured in config.json. The file experiments_done.db is a SQLite database that contains the experiments already done. The requeriments.txt file contains the required packages to run the experiments. The Python version used is 3.10.8.

Fundings

This work was supported through the Junta de Castilla y León (JCyL) (regional government) under project BU055P20 (JCyL/FEDER, UE), the Spanish Ministry of Science and Innovation under project PID2020-119894GB-I00 co-financed through European Union FEDER funds, and project TED2021-129485B-C43 funded by MCIN/AEI/10.13039/501100011033 and the European Union NextGenerationEU/PRTR. J.L. Garrido-Labrador is supported through Consejería de Educación of the Junta de Castilla y León and the European Social Fund through a pre-doctoral grant EDU/875/2021 (Spain).