Jadwal Sholat

Memuat jadwal sholat…

Ilmu Komputer & AI editorial

Open AccessOA2024

Hybrid Approaches to Optimization and Machine Learning Methods: A Systematic Literature Review

This systematic literature review analyzes hybrid methods that integrate optimization and machine learning techniques to overcome individual limitations in clustering and classification. It provides a bibliometric overview, state-of-the-art review, and SWOT analysis of ten most cited algorithms, identifying opportunities for hybrid approaches.
Beatriz Flamia Azevedo; Ana Maria A. C. Rocha; Ana I. Pereira· Machine-mediated learning· 2024· DOI 10.1007/s10994-023-06467-x

The core problem

Real-world problems are increasingly complex, requiring sophisticated models and algorithms that can efficiently handle large datasets and find optimal solutions. However, no single method or algorithm is perfect; each has limitations that can be mitigated or eliminated by combining different methodologies. Hybrid algorithms are expected to leverage the strengths of both optimization and machine learning to create more efficient integrated methodologies. This paper presents an extensive systematic and bibliometric literature review on hybrid methods involving optimization and machine learning techniques for clustering and classification. The aim is to identify the potential of methods and algorithms to overcome the difficulties of one or both methodologies when combined. After describing optimization and machine learning methods, a numerical overview of works published since 1970 is presented, followed by an in-depth state-of-the-art review over the last three years. A SWOT analysis of the ten most cited algorithms from the collected database investigates the strengths and weaknesses of pure algorithms and highlights the opportunities and threats explored with hybrid methods. This investigation highlights notable works and discoveries involving hybrid methods in clustering and classification and points out the difficulties of pure methods that can be strengthened through inspirations from other methodologies—hybrid methods.

Innovation

The review methodology follows a systematic approach: a comprehensive search of the literature on hybrid methods combining optimization and machine learning for clustering and classification was conducted. The authors performed a bibliometric analysis to provide a numerical overview of publications since 1970, identifying trends and key contributors. Subsequently, an in-depth state-of-the-art review focused on the last three years was conducted to capture recent advances. A SWOT analysis was performed on the ten most cited algorithms from the collected database, examining their strengths and weaknesses in pure form and the opportunities and threats that arise when hybridized. This structured approach allows for a rigorous assessment of the current landscape and future directions of hybrid methods.
Introduction
Real-world problems are increasingly complex, requiring sophisticated models and algorithms that can efficiently handle large datasets and find optimal solutions. However, no single method or algorithm is perfect; each has limitations that can be mitigated or eliminated by combining different methodologies. Hybrid algorithms are expected to leverage the strengths of both optimization and machine learning to create more efficient integrated methodologies. This paper presents an extensive systematic and bibliometric literature review on hybrid methods involving optimization and machine learning techniques for clustering and classification. The aim is to identify the potential of methods and algorithms to overcome the difficulties of one or both methodologies when combined. After describing optimization and machine learning methods, a numerical overview of works published since 1970 is presented, followed by an in-depth state-of-the-art review over the last three years. A SWOT analysis of the ten most cited algorithms from the collected database investigates the strengths and weaknesses of pure algorithms and highlights the opportunities and threats explored with hybrid methods. This investigation highlights notable works and discoveries involving hybrid methods in clustering and classification and points out the difficulties of pure methods that can be strengthened through inspirations from other methodologies—hybrid methods.

Why it matters

The review underscores the potential of hybrid methods to overcome the limitations of individual optimization and machine learning techniques. By combining methodologies, hybrid algorithms can achieve better performance in terms of solution quality, convergence speed, and robustness. However, the design of effective hybrid methods requires careful consideration of the balance between exploration and exploitation, computational cost, and problem-specific characteristics. The SWOT analysis provides a framework for understanding when and how to hybridize, guiding researchers and practitioners. Future research directions include developing adaptive hybrid strategies, integrating deep learning with optimization, and addressing scalability for big data. The authors conclude that hybrid methods are a promising avenue for tackling complex real-world problems in clustering and classification.

Who should read this

Pembaca Research Digest dan praktisi rumpun ini.

Opening member content…