ClusterValidityIndices.jl
ap6yc.github.io
- Free plan
- Linux
7.1 Club score Visit ap6yc.github.io
What the maker publishes
- Batch and incremental modes
- The package implements both batch and incremental variants, but one CVI object cannot switch between those modes. source
- Batch and incremental use
- Its indices have batch and incremental implementations for evaluating clustering results as a batch or as samples arrive. source
- Data format
- The package expects floating-point feature matrices with samples in columns and integer cluster labels. source
- Data import formats
- Float matrices with integer cluster-label vectors source
- Deployment
- desktop source
- Distribution
- The project is distributed as a Julia package through JuliaHub and can also be installed directly from GitHub. source
- Documentation build
- The official documentation reports that it was built with Julia 1.8.5 on Linux on January 13, 2023. source
- Examples
- The repository says its example scripts can be opened, saved, and run as Julia notebooks. source
- Free plan
- Yes source
- Implemented indices
- The documented indices are Calinski-Harabasz, centroid-based Silhouette, Davies-Bouldin, Generalized Dunn 43 and 53, Partition Separation, representative Cross Information Potential, WB-index, and Xie-Beni. source
- Included indices
- The documented indices include Calinski-Harabasz, centroid-based Silhouette, Davies-Bouldin, two generalized Dunn indices, Partition Separation, representative Cross Information Potential, WB-index, and Xie-Beni. source
- Incremental evaluation
- Incremental cluster validity indices support evaluating a clustering process online and are numerically equivalent to their batch counterparts. source
- Input format
- The guide expects floating-point data matrices with samples in columns and features in rows, plus integer cluster labels. source
- Installation
- The package can be installed through Julia's package manager or directly from its GitHub repository. source
- Intended users
- The package is intended for evaluating unsupervised clustering algorithms using data samples and algorithm-assigned cluster labels. source
- License
- The repository states that the software is maintained under the MIT License. source
- Maintainer
- The contribution guide names Sasha Petrenko as the current developer and maintainer. source
- Mode limitation
- A CVI object cannot switch between batch and incremental evaluation; a fresh object is needed to use the other mode. source
- Purpose
- The package provides cluster validity indices to assess unsupervised clustering when ground-truth labels are unavailable. source
- Scripting support
- Yes source
- Security and compliance
- The reviewed project pages state an MIT license but do not describe security certifications or compliance standards. source
- Support
- The contribution guide identifies GitHub Issues as the main contact point for questions, concerns, and requested changes. source
- Support
- The project directs questions and requested changes to its public GitHub issues page. source
Checked against the maker's pages on 12 Oct 2026.