Bio-inspired algorithms for feature engineering: analysis, applications and future research directions
| Date | 18 April 2024 |
| Pages | 56-71 |
| DOI | https://doi.org/10.1108/IDD-11-2022-0118 |
| Published date | 18 April 2024 |
| Subject Matter | Library & information science,Library & information services,Lending,Document delivery,Collection building & management,Stock revision,Consortia |
| Author | Vaishali Rajput,Preeti Mulay,Chandrashekhar Madhavrao Mahajan |
Bio-inspired algorithms for feature
engineering: analysis, applications and
future research directions
Vaishali Rajput and Preeti Mulay
Department of Engineering, Symbiosis International University, Pune, India, and
Chandrashekhar Madhavrao Mahajan
Department of Engineering Sciences and Humanities, Vishwakarma Institute of Technology, Pune, India
Abstract
Purpose –Nature’s evolution has shaped intelligent behaviors in creatures like insects and birds, inspirin g the field of Swarm Intelligence.
Researchers have developed bio-inspired algorithms to address complex optimization problems efficiently. These al gorithms strike a balance
between computational efficiency and solution optimality, attracting significant attention across domains.
Design/methodology/approach –Bio-inspired optimization techniques for feature engineering and its applications are systematical ly reviewed
with chief objective of assessing statistical influence and significance of “Bio-inspired optimization”-based computational models by referring to
vast research literature published between year 2015 and 2022.
Findings –The Scopus and Web of Science databases were explored for review with focus on parameters such as country-wise publications,
keyword occurrences and citations per year. Springer and IEEE emerge as the most creative publishers, with indicative prominent and superior
journals, namely, PLoS ONE, Neural Computing and Applications, Lecture Notes in Computer Science and IEEE Transactions. The “National Natural
Science Foundation”of China and the “Ministry of Electronics and Information Technology”of India lead in funding projects in this area. China,
India and Germany stand out as leaders in publications related to bio-inspired algorithms for feature engineering research.
Originality/value –The review findings integrate various bio-inspired algorithm selection techniques over a diverse spectrum of opt imization
techniques. Anti colony optimization contributes to decentralized and cooperative search strategies, be e colony optimization (BCO) improves
collaborative decision-making, particle swarm optimization leads to exploration-exploitation balance and bio-inspired algorithms offer a rangeof
nature-inspired heuristics.
Keywords Bio-inspiredalgorithms, Evolutionaryalgorithms, Speech emotionrecognition, Sentiment analysis,Nature-inspiredoptimization algorithms,
Feature engineering, Bio-inspiredcomputing, Bio-inspired optimization, Bio-inspiredapplication, Feature selection
Paper type Literature review
1. Introduction
Nature has molded cognitive behaviors and biological
phenomena over millions of years through the process of
evolution, giving animals like insects and birds their flexibility,
self-learning and proficiency. The systematic study of
intelligent evolutionary behaviors of insect’s/organism’s
structure, functioning and interrelationships have gained an
imperative importance in view of its usability in modern
artificially intelligent technologies. Social behaviors in ant
colonies (Shokouhifar,2011;Dorigo et al.,2006;Amarjeetand
Chhabra, 2017), beehives (bee colony optimization [BCO])
(Todorovics, 2009) and bird flocks, have given rise to swarm
intelligence (Del Ser et al.,2019;Christo et al.,2019;Darwish,
2018). Researchers in the computational community have
created methods to solve challenging modeling, simulation and
optimizationissues by drawing inspiration from these biological
systems. Optimization problemsfrom a variety of domains also
have attracted significant attention, leading to the creation of
bio-inspired investigative approaches that strike a balance
between computational efficiency and solution optimality
(Christo et al.,2019;Wang et al.,2018). Publications about
statistical data from bio-inspired algorithms have multiplied
dramatically in the last several years. Reasonable convergence
rate, low processing time,exploration–exploitation balance and
few algorithm-specific control parameters are the essential
components of a moral bio-inspired algorithm (Christo et al.,
2019;Darwish, 2018). Convergence implies that an algorithm
is able to approach the global or nearlyideal answer quickly.
Faster convergence refers to the algorithm’s ability to arrive at a
good solution with fewer iterations or solution assessments.
Fundamentally, each bio-inspired algorithm consists of two parts:
exploitation (improving upon existing solutions) and exploration
(discovering novel techniques for solutions). When tackling
The current issue and full text archiveof this journal is available on Emerald
Insight at: https://www.emerald.com/insight/2398-6247.htm
Information Discovery and Delivery
53/1 (2025) 56–71
© Emerald Publishing Limited [ISSN 2398-6247]
[DOI 10.1108/IDD-11-2022-0118]
Received 19 November 2022
29 April 2023
23 October 2023
7 January 2024
3 February 2024
Accepted 17 March 2024
56
optimization problems, finding the ideal balance between
exploration and exploitation is essential. Overemphasis on either
exploration or exploitation can lead to suboptimal problem
solutions. A worthy algorithm adapts its behavior to the nature
and specificity of problem’s characteristics, confirming a rational
and sensible approach. Exploration guarantees that the search
space has a range of favorable regions that the algorithm can find,
and exploitation guarantees that those regions include the best
solutions. The best possible outcome for a particular situation
requires the adjustment of these characteristics. An algorithm that
uses the least number of tunable parameters is required due to its
competence to reduce the algorithm design complexity and fine-
tuning. Complex algorithms with numerous control parameters
always face optimization challenges and requires extensive tuning
of parameters, leading to less practical, sluggish to solve real-wor ld
problems and applications.
The feature engineering research field that requires audio
data analysis, such as speech emotion recognition (SER), has
recently gainedan importance as evidentfrom the large number
of publications. Extracting relevant and discriminative features
of audio signals to recognize the subject’s emotional state is a
challenging task. So far, researchers have explored several low-
level and high-quality features for SER, which include “zero-
crossing rate,”“Mel-Frequency Cepstral Coefficients
(MFCCs)”spectral features and features related with pitch. In
the current scenario, scientists are confrontingsignificant issues
in feature engineering, such as how to choose a resilientstrategy
to separate predominant and selective parameters from audio
waves to address the enthusiastic condition of a person from
their sound transmission. In the previous era, numerous
specialists have examined low-level high-quality features for
SER such as zero-intersection, straight indicator coefficient,
pitch, energy,Mel-recurrence MFCC and nonlinear highlights,
for example,tiger energy administrator.
In an effort to find workable, high-quality solutions for novel,
challenging optimization issues, researchers began using heuristics.
Heuristics has emerged as a worthy approach for finding feasible,
finest and superior solutions to complex optimization problems.
Heuristic problem-solving techniques, though do not promise
optimal solutions, are designed and devised to obtain reliable
solutions in rational time spans (Del Ser et al., 2019;Christo et al.,
2019;Darwish, 2018). In the realm of machine learning, a
metaheuristic algorithm is created to identify the best answers to
particular complex situations. Some of these algorithms take their
cues from the way animals or birds behave in the wild.
To achieve the optimized solution,metaheuristic algorithms use
a function called heuristic (Del Ser et al., 2019;Christo et al.,
2019;Darwish, 2018). Researchers constantly endeavor to
advance feature engineering techniques and develop robust tactics
and strategies to extract predominant and selective parameters in
audio signals for enhanced accuracy and robustness while
recognizing the dynamic and real-world emotional envir onments.
The common issues and challenges like intractability of complex
problems, efficiency, adaptability, scalability, trade-offs, multi-
objective optimization and problem-specific customization can be
effectively addressed and solved with heuristic approaches.
Problems related to “swarm optimization”(Wang et al., 2018;
Ahmad, 2015;Nguyen et al., 2020;Lin et al.,2016), “hybrid
swarm optimization”(Ampellioa nd Vassio, 2016), “binaryswarm
optimization”(Hızarcıet al., 2022;Abualigah et al., 2019)and
“evolutionary algorithms”(Zhao and Jin-Hu, 2015;Abroudi et al.,
2013) can be well addressed and optimized with various heuristic
and metaheuristic approaches. The heuristic approaches and
techniques leverage nature-inspired and driven ideas, physics
principles and social behavior traits and characteristics to steer and
guide toward the search for reasonably acceptable working
solutions (Kumar and Singh, 2021;Sreejith et al.,2020).
These days, researchers use deep learning approaches to
restrain perception issues, for example, voice identification,
feeling identification, signal identification, face identification
and image identification (Kumar and Singh, 2021;Rodrigues
et al., 2015;Rani and Ramyachitra,2017;Gandomi and Alavi,
2012;Alweshah et al., 2022).In addition, researchers have also
explored various other advanced techniques for SER, which
includes deep learningmodels that automatically learns directly
from raw audio data to estimate the discriminative features
(Sreejith et al., 2020). The benefit of using the profound
learning approaches relies upon the involuntary feature
formation, selection and extraction in such a way that the
recommended model tracks down the significant and the low-
level characteristics through the adaptableconvolution weights
concerning the input information (Anter et al.,2015;Abd-
Alsabour, 2016;Inamdar et al., 2021). Research on simulating
natural behaviors to solve complicated computational issues
has significantly increased in the past few years, particularly in
the optimizationareas. Despite this growth, certain areas within
bio-inspired optimizationrequire more exploration. This paper
outlines the current state of the field, identifiesopen challenges
and emphasizes the need for collaboration to gain valuable
insights into these optimizationtechniques. These approaches,
habitually referred to as bio-inspired or nature-inspired
optimizations, extract motivation from biological and natural
systems to design algorithms and heuristics. Some examples
include “genetic algorithms,”“ant colony optimization,”
“particle swarm optimization”and “simulated annealing.”
Though, regardless of these developments and frequent
successes in applying bio-inspired techniques to various
problems, there are several areas within this field that warrant
further exploration and development. Some important aspects
that require continual attention that researchers are trying to
address include scalability, robustness, hybridization or
combination of different optimization techniques, dynamic
environment adaptability, multiobjectiveness, real-world
applications, theoretical frameworks and foundations and
parallel and distributedcomputing environments.
Thus, the dynamic and evolving nature of bio-inspired or
nature-driven optimization field is steering researchers to
address various challenges to explore new frontiers to improve
the efficiency, robustness and applicability for solving a wide
range of complex computationalproblems. This review paper’s
primary goal is to present a thorough examination of several
bio-inspired feature engineering algorithms. This paper also
aims to estimate the importance of meta-heuristic algorithms
by analyzing the research articles published between 2015 and
2022. Data is often believed to be the basis for decision-
making, innovation and business success; hence, many experts
work hard to retrieve crucial data in the competitive era. This
work investigates, reviews, compares, analyses and provides
perceptions on a variety of literature reports on feature
engineering aspects of bio-inspired optimization techniques in
Bio-inspired algorithms
Vaishali Rajput, Preeti Mulay and Chandrashekhar Madhavrao Mahajan
Information Discovery and Delivery
Volume 53 · Number 1 · 2025 · 56–71
57
Get this document and AI-powered insights with a free trial of vLex and Vincent AI
Get Started for FreeStart Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
