IntelELM: A python framework for intelligent metaheuristic-based extreme learning machine

This study introduces IntelELM, an open-source Python library designed for hybrid neural networks that integrate Extreme Learning Machine (ELM) with Metaheuristic Algorithms (MHAs). Built on the foundations of two well-established libraries, Scikit-Learn and Mealpy, IntelELM offers four primary stra...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Neurocomputing (Amsterdam) Ročník 618; s. 129062
Hlavní autoři: Van Thieu, Nguyen, Houssein, Essam H., Oliva, Diego, Hung, Nguyen Duy
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 14.02.2025
Témata:
Extreme learning machine
Machine learning
Metaheuristic algorithms
Metaheuristic optimization-based ELM
Neural network
Python library
Neural network
Extreme learning machine
Python library
Metaheuristic optimization-based ELM
Machine learning
Metaheuristic algorithms
ISSN:0925-2312
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:This study introduces IntelELM, an open-source Python library designed for hybrid neural networks that integrate Extreme Learning Machine (ELM) with Metaheuristic Algorithms (MHAs). Built on the foundations of two well-established libraries, Scikit-Learn and Mealpy, IntelELM offers four primary strategies for addressing regression and classification tasks. These strategies are implemented through the ElmRegressor and ElmClassifier classes for traditional ELM, as well as the MhaElmRegressor and MhaElmClassifier for hybrid metaheuristic-based ELM models. The library is easy to install and use, especially for individuals familiar with the Scikit-Learn ecosystem. IntelELM comprises at least 402 distinct models across these four primary classes, encompassing classical ELM regression and classification models, as well as over 200 metaheuristic-based ELM regression and classification models each. To demontrade the power of the proposed library, we evaluate several hybrid models from the IntelELM library alongside traditional machine learning models across three benchmark datasets. Experimental results demonstrate that the hybrid models within IntelELM exhibit competitive performance across various performance metrics compared to traditional machine learning approaches. These findings underscore the library's potential to offer effective solutions to real-world problems and contribute significantly to the computer science community. We have released the source code of the library as open-source, inviting the research community to conduct widespread evaluations of this comprehensive framework as a promising tool for research studies and real-world solutions. The source code can be found at https://github.com/thieu1995/IntelELM. •We propose hybrid-based Extreme Learning Machine (ELM) software called IntelELM.•IntelELM provides classical and metaheuristic-based ELM models for diverse tasks.•Easy installation and use for those familiar with the Scikit-Learn ecosystem.•Hybrid models in IntelELM show competitive performance on benchmark datasets.•IntelELM's potential for effective real-world problems and ML community impact.
ISSN:0925-2312
DOI:10.1016/j.neucom.2024.129062