GitHub - mike-liuliu/gl_index: Source code of the paper "A New Index for Clustering Evaluation Based on Density Estimation."

mike-liuliu / gl_index Public

Notifications You must be signed in to change notification settings
Fork 0
Star 2

Source code of the paper "A New Index for Clustering Evaluation Based on Density Estimation."

arxiv.org/abs/2207.01294

Apache-2.0 license

2 stars 0 forks Branches Tags Activity

Star

Notifications

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
data		data
funcs		funcs
results of all 145 data		results of all 145 data
.DS_Store		.DS_Store
.gitattributes		.gitattributes
Data sources of the 145 datasets.txt		Data sources of the 145 datasets.txt
LICENSE.txt		LICENSE.txt
README.txt		README.txt
calculate gl_index.ipynb		calculate gl_index.ipynb
requirements.txt		requirements.txt
test CDbw.ipynb		test CDbw.ipynb
test DBCV.ipynb		test DBCV.ipynb
test Silhouette coefficient (SC).ipynb		test Silhouette coefficient (SC).ipynb
test VIASCKDE.ipynb		test VIASCKDE.ipynb
test calinski_harabasz_index (CH).ipynb		test calinski_harabasz_index (CH).ipynb
test davies_bouldin_index (DB).ipynb		test davies_bouldin_index (DB).ipynb
test gl_index - Bayesian Gaussian Mixture version.ipynb		test gl_index - Bayesian Gaussian Mixture version.ipynb
test gl_index.ipynb		test gl_index.ipynb

Repository files navigation

0. This is the source code of the paper "A New Index for Clustering Evaluation Based on Density Estimation."

1. In function index_plot_first_n_label_one_data, if the index's score is "smaller is better", then the "smaller_better" hyper-parameter should be set to True. Otherwise, if the index's score is "larger is better", then the "smaller_better" hyper-parameter should be set to False.

2. Readers can test their own index function, the API is:
   
    def index_function(X, label):
   
    some codes to compute the index value ...
    
    return the_index_value

then call the index_plot_first_n_label_one_data function. Note the "smaller_better" hyper-parameter.

3. License.
   
License of the source code : Apache License, Version 2.0
License of new data: Creative Commons Attribution 4.0 International

4. Citation:

@article{liu2022new,
  title={A new index for clustering evaluation based on density estimation},
  author={Liu, Gangli},
  journal={arXiv preprint arXiv:2207.01294},
  year={2022}
}

5. The "multiple_label_145.p" file is larger than 100MB, so it was compressed to make it smaller. Readers need to unzip the "multiple_label_145.p.zip" file in the "data" folder.