https://github.com/aclai-lab/solefeatures.jl

Perform feature selection on unstructured and multimodal data!

https://github.com/aclai-lab/solefeatures.jl

Science Score: 36.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Committers with academic emails
    1 of 9 committers (11.1%) from academic institutions
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (14.2%) to scientific vocabulary

Keywords from Contributors

decision-trees logic modal-logic symbolic-learning time-series-classification
Last synced: 10 months ago · JSON representation

Repository

Perform feature selection on unstructured and multimodal data!

Basic Info
  • Host: GitHub
  • Owner: aclai-lab
  • License: mit
  • Language: Julia
  • Default Branch: main
  • Homepage:
  • Size: 352 KB
Statistics
  • Stars: 7
  • Watchers: 4
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 4 years ago · Last pushed about 1 year ago
Metadata Files
Readme License

README.md

SoleFeatures.jl – Feature Selection on Unstructured and Multimodal data

Stable Build Status Coverage Code Style: Blue

In a nutshell

SoleFeatures.jl provides tools for filter-based feature selection on (un)structured data. At this time, the package provides: - 3 native feature selection methods, plus a wrapper around python implementations from scikit-learn and scikit-feature; - Generalized feature selection methods that also apply to dimensional data (e.g., images or time-series), via a step of window-based flattening; - Specific methods for time-series feature selection, based on Catch22.jl; - An easily extendible codebase, with abstraction layers similar to those of scikit-learn.

Outside the nutshell

In machine learning, feature selection is the practice of reducing the size of a dataset by selecting only a subset of relevant features. This reduces the time that is necessary for learning models, and it generally leads to models that are more performant (in fact, irrelevant data often misleads machine learning algorithms). Despite the increasing importance of this data processing practice, not only there is not a proper package for it in the Julia environment, but there is also a lack of implemented algorithms for it.

So, here comes SoleFeatures.jl, which offers a simple interface for performing feature selection with minimal effort on dataset structures implementing the Table.jl interface. The package also allows to access partial results of the feature selection processes (e.g., feature importance scores), offering complete control over the whole process. At this time, the package provides 3 native methods based on variance, correlation and statistical hypothesis testing, as well as a wrapper unlocking the majority of filter-based methods available in the python libraries scikit-learn and scikit-feature.

These classical feature selection methods are designed for (un)supervised structured datasets, that is, tabular datasets with scalar features; however, an interesting characteristic of this package is the possibility of using these methods on unstructured data as well, for example, with time-series and images. This generalization is possible via a prior window-based flattening step, where tailored measures are used to compress the non-scalar components of the data. For example, with time-series data, measures with solid statistical grounds from Catch22.jl are used.

All feature selection methods can also be applied to multimodal datasets as well. Finally, the package offers an easily extendible codebase, with abstraction layers similar to those of scikit-learn, and was built as part of Sole.jl, a novel programming suite tailored for unstructured symbolic learning.

About

The package is developed by the ACLAI Lab @ University of Ferrara.

SoleFeatures.jl was built for Sole.jl, an open-source framework for symbolic machine learning.

Owner

  • Name: Applied Computational Logic and Artificial Intelligence Laboratory
  • Login: aclai-lab
  • Kind: organization
  • Email: aclai@unife.it
  • Location: Italy

Applied Computational Logic and Artificial Intelligence (ACLAI) Laboratory of the Department of Mathematics and Computer Science, University of Ferrara

GitHub Events

Total
  • Watch event: 2
  • Push event: 28
  • Pull request event: 8
  • Create event: 1
Last Year
  • Watch event: 2
  • Push event: 28
  • Pull request event: 8
  • Create event: 1

Committers

Last synced: over 1 year ago

All Time
  • Total Commits: 136
  • Total Committers: 9
  • Avg Commits per committer: 15.111
  • Development Distribution Score (DDS): 0.382
Past Year
  • Commits: 2
  • Committers: 2
  • Avg Commits per committer: 1.0
  • Development Distribution Score (DDS): 0.5
Top Committers
Name Email Commits
PatrikCavina p****a@e****t 84
giopaglia 2****a 19
PatrikCavina c****k@g****m 14
PasoStudio73 p****3@g****m 10
ferdiu f****a@g****m 3
eduardstan s****d@g****m 2
CompatHelper Julia c****y@j****g 2
Michele21 g****2@g****m 1
Alberto Paparella 5****a 1
Committer Domains (Top 20 + Academic)

Issues and Pull Requests

Last synced: 11 months ago

All Time
  • Total issues: 0
  • Total pull requests: 42
  • Average time to close issues: N/A
  • Average time to close pull requests: 7 months
  • Total issue authors: 0
  • Total pull request authors: 1
  • Average comments per issue: 0
  • Average comments per pull request: 0.0
  • Merged pull requests: 2
  • Bot issues: 0
  • Bot pull requests: 42
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Top Authors
Issue Authors
Pull Request Authors
  • github-actions[bot] (36)
Top Labels
Issue Labels
Pull Request Labels

Dependencies

.github/workflows/TagBot.yml actions
  • JuliaRegistries/TagBot v1 composite