A newly proposed algorithmic framework designated as Self-Adaptive Improved Fungal Growth Optimization consistently outperforms or remains highly competitive with multiple existing state-of-the-art metaheuristic algorithms across diverse datasets, according to experimental findings published in Knowledge and Information Systems. The research introduces this computational approach to deliver high-quality, robust, and scalable clustering solutions for complex data clustering problems.
Within the broader domains of machine learning and data science, data clustering represents a fundamental computational task. However, established and traditional clustering algorithms frequently exhibit critical operational weaknesses. In particular, conventional methods remain highly sensitive to initialization values and are often prone to becoming trapped in local optima, a limitation that is especially pronounced when algorithms process complex or high-dimensional datasets.
To overcome these longstanding operational limitations, the study proposes the Self-Adaptive Improved Fungal Growth Optimization, abbreviated as SAIFGO. The method is developed as an enhanced version of the nature-inspired Fungal Growth Optimization algorithm. Computationally, the proposed SAIFGO framework mimics key biological fungal behaviors, specifically simulating hyphal growth, chemotropism, branching, and spore germination during the search process.
To systematically improve the optimization and search process, the authors integrated several advanced computational strategies directly into SAIFGO. The framework incorporates chaotic initialization to establish initial configurations, alongside Lévy flight-based mutation to aid exploration. Furthermore, the algorithm implements opposition-based learning and utilizes an adaptive exploration–exploitation mechanism designed to improve the progression of the search.
To achieve superior clustering performance, the optimization framework employs a composite multi-objective fitness function. This evaluation function is formulated using multiple cluster validity indices. In addition, the proposed approach is hybridized with conventional clustering techniques, specifically incorporating K-Means and Fuzzy C-Means, which allows the hybrid framework to optimize cluster centroids more effectively.
The researchers conducted extensive experimental evaluations by comparing the proposed SAIFGO framework against several contemporary state-of-the-art metaheuristic optimization algorithms. These comparative experiments evaluated performance using five distinct cluster validity indices alongside comprehensive statistical analyses across diverse datasets to assess the consistency, effectiveness, and scalability of the resulting cluster solutions.
The experimental results demonstrate that the proposed SAIFGO algorithm consistently outperforms or remains highly competitive with existing optimization methods across the tested datasets. Based on these statistical and experimental analyses, the authors conclude that SAIFGO serves as an effective, robust, scalable, and high-performing optimization algorithm specifically tailored for solving complex data clustering problems.
Subject of Research: Metaheuristic optimization algorithms for data clustering
Article Title: Self-adaptive improved fungal growth optimization: a hybrid metaheuristic framework for high-quality, robust, and scalable clustering
Article References: Das, K. K., Sarkar, S., Majumdar, A., & Mukhopadhyay, A. (2026). Self-adaptive improved fungal growth optimization: a hybrid metaheuristic framework for high-quality, robust, and scalable clustering. Knowledge and Information Systems, 68(1), Article 317. https://doi.org/10.1007/s10115-026-02935-7
Image Credits: AI Generated
DOI: 10.1007/s10115-026-02935-7
Keywords: data clustering, machine learning, data science, SAIFGO, fungal growth optimization, metaheuristic algorithms, hyphal growth, chemotropism, spore germination, chaotic initialization, Lévy flight mutation, opposition-based learning, cluster validity indices, K-Means, Fuzzy C-Means, local optima, cluster centroids
Cite Scienmag News
Blake Davidson. (November 17, 2026). Fungal-Inspired Optimization Algorithm Outperforms Existing Methods in Complex Data Clustering. Scienmag. https://scienmag.com/fungal-inspired-optimization-algorithm-outperforms-existing-methods-in-complex-data-clustering/
Blake Davidson. "Fungal-Inspired Optimization Algorithm Outperforms Existing Methods in Complex Data Clustering." Scienmag, 17 November 2026, https://scienmag.com/fungal-inspired-optimization-algorithm-outperforms-existing-methods-in-complex-data-clustering/. Accessed 17 November 2026.
Blake Davidson. "Fungal-Inspired Optimization Algorithm Outperforms Existing Methods in Complex Data Clustering." Scienmag. November 17, 2026. https://scienmag.com/fungal-inspired-optimization-algorithm-outperforms-existing-methods-in-complex-data-clustering/








