Full length articleModified Grasshopper Optimization Algorithm for detection of Autism Spectrum Disorder
Introduction
Autism spectrum disorder (ASD) is a neurological disorder characterized by weak cognitive skills and other heterogeneous behavioral symptoms like social impairment, resistance to change, and indistinct speech in some cases [1], [2]. ASD patients often experience communication abnormalities like absenteeism of language or delayed language. In contrast to other characteristics of ASD, the ability to socialize is greatly affected. It diversifies with age, IQ, and the environment. One of the most prevalent disorders of development is autism. According to the statistics, ASD impacts more than 1% of the inhabitants, and males are quadruple times more susceptive than the opposite gender [3]. A recent report by the Centers for Disease Control (CDC) estimates the ubiquity of ASD in the United States as 1 in 68, a substantial increase in the last decade [4].
Autism spectrum disorders prevail in various forms ranging from extremely mild ASD to severe ASD contingent on the severity of the occurring symptoms. The symptoms may occur during early childhood, adolescence, or adulthood. Early detection of mild forms of ASD is quite difficult in particular because several diagnoses depict a range of common neurodevelopmental symptoms [5]. Currently, autism stays unidentified until four years of age, frequently after more than two years of medical consultation [6]. The lag in the diagnosis of autism postpones the required autism-specific intervention. Therefore, early diagnosis and treatment of ASD are the need of the hour.
Contemporary ASD detection techniques broadly rely on the particular domain expert and the various questions as answered by the individuals under examination. But the scholars have criticized these traditional methods for being subjective and tedious [7], [8], [9], [10], [11]. The union of machine learning and the biomedical domain has been profusely beneficial in detecting several diseases such as Cancer, Diabetes, Celiac, Alzheimer, Parkinson’s, etc. Since ASD is ubiquitous and the complete cure of the disease is not possible to date, we decided to work on ASD, whose early diagnosis has always been considerably challenging.
This research paper works on the stochastic detection of autism among children, adolescents, and adults using the proposed Modified Grasshopper Optimization Algorithm (MGOA). Grasshopper Optimization Algorithm (GOA) is a nature-inspired algorithm mainly based on exploration and exploitation characteristics of swarms. We have purposely employed GOA for feature selection, as a result of some of the existing works we came across, for instance, Houssein et al. [12] identified the ability of GOA to resolve real-world problems having anonymous search space. They used GOA with Support Vector Machines for the automatic detection of seizures in EEG. Their empirical results demonstrated that the proposed GOA+SVM approach is capable of detecting epilepsy and enhancing the diagnosis with 100% accuracy for standard data versus epileptic data. Tumuluru et al. [13] proposed the Chronological Grasshopper Optimization Algorithm (Chronological-GOA) for collecting gene information and classification of cancer. For identification, the proposed Chronological GOA algorithm selects the data needed to be sent to the Deep Belief Network (DBN). The accuracy of the proposed DBN Chronological-GOA is 0.9767 for the colon database and 0.9729 for the leukemia database. Faris et al. [14] introduced a new stochastic hybrid learning method using the theoretical grasshopper optimization algorithm (GOA) for neural networks with multilayer perceptrons (MLP). The suggested GOAMLP framework is then extended to five datasets following the introduction of an encoding method and objective feature based on mean-squared error (MSE): breast cancer, Parkinson, hypertension, coronary heart disease, and orthopedic patients. It has been depicted that the proposed GOAMLP stochastic learning is very useful in enhancing the MLP classification level. Taking motivation from such existing works, we were deterministic to choose GOA for the detection of ASD as a result of its evident robustness and potential to derive highly effective results
Through this paper, we intend to contribute to the society by reducing the period of the diagnostic odyssey that the affected families go through and expand the scope of the patients to benefit from early access to diagnostic intervention.
The proposed paper features the following:
Grasshopper optimization algorithm has been employed for efficient feature extraction.
Random forest is used as a classifier for the evaluation of the quality of the extracted features.
Modified GOA is applied to three ASD screening datasets, mentioned further in the paper, targeting all age groups namely, children, adolescents, and adults, for the prediction of ASD with an approximate accuracy of 100%.
The objective of the paper is to help the sufferers get early access to the accurate diagnosis of ASD. Hence, making our contribution to the community.
The paper is divided into following sections: Section 2 gives the details of the traditional GOA, Section 3 provides the proposed algorithm, i.e. Modified Grasshopper Optimization Algorithm, Section 4 provides the information about the implementation of the proposed algorithm, it includes the experimental setup, input parameters and the description of the datasets considered. Section 5 presents the results obtained on all the datasets considered by applying the proposed algorithm. Section 6 provides the comparative study of the proposed work to the existing algorithms. Section 7 of the paper states the conclusion drawn from the results obtained, and the last section provides references to the work of different related authors.
Section snippets
Existing works on autism spectrum disorder
In recent years, research has grasped various heuristic and statistical models to comprehend and analyze the methods to diagnose and cure autism spectrum disorders. Machine learning is pertinent to recognize complex paradigms [15]. Therefore, machine learning algorithms are used to execute the binomial classification task of identifying features for the detection of the disorder. Some studies focused on autism research. Specifically, the diagnosis of ASD [16], [17], [18], [19], [20], [21],
Proposed grasshopper optimization algorithm
This section presents the proposed Modified Grasshopper Optimization Algorithm. The proposed algorithm has been implemented in python using its various libraries. The proposed algorithm has been applied for the detection of widespread Autism Spectrum Disorders (ASD) in all the stages of life. The algorithm proposed is described below:
Algorithm 2. Modified Grasshopper Optimization Algorithm
Fig. 4 shows the flowchart for the proposed Modified Grasshopper Optimization Algorithm. We have
Implementation
This section of the paper discusses the experimental setup, input parameters and description of the datasets considered for the implementation of the proposed algorithm.
Results and discussion
This section of the proposed paper discusses the results obtained when the proposed Modified Grasshopper Optimization Algorithm has been applied to all the datasets of Autism Spectrum Disorder. Various performance metrics have been used to measure the performance of the proposed algorithm for each dataset.
When Modified Grasshopper Optimization Algorithm was implemented, the following Table 3, Table 4, Table 5 represent the results of each dataset. For each dataset, we considered the fitness
Comparison of proposed work with existing algorithms
This section of the paper compares the results obtained using the proposed algorithm with the earlier proposed works for the detection of Autism Spectrum Disorder.
We have proposed an algorithm to detect Autism Spectrum Disorders, and our primary concern was to overcome the barrier of age. The proposed algorithm has been compared with traditional GOA as well as 5 different existing algorithms. The MGOA has outperformed well in three of the dataset as shown in Fig. 11 and Table 7. The proposed
Conclusion and future work
This research work focuses on the early diagnosis of Autism Spectrum Disorder which can save lives across the globe using the proposed Modified Grasshopper Optimization Algorithm. Delayed diagnosis of a disease is responsible for the deterioration of the quality of life of the sufferer and the associated people. Early as well as accurate detection is of paramount importance. Feature Selection plays a vital role in developing a model. It helps to eliminate the least impact creating features. For
CRediT authorship contribution statement
Nikita Goel: Conceptualization, Data curation. Bhavya Grover: Formal analysis, Funding acquisition. Anuj: Investigation, Methodology. Deepak Gupta: Supervision, Validation, Visualization. Ashish Khanna: Project administration, Resources, Software. Moolchand Sharma: Writing - original draft, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Miss. Nikita Goel was born in India (1997) and currently residing in Delhi. She is pursuing her B.Tech. in Computer Science and Engineering from Maharaja Agrasen Institute of Technology (2016–2020), GGSIPU, Delhi. She has been conferred on as the Senior Under Officer in NCC. She has gained expertise in Python, Machine Learning and is improving her Deep Learning skills in her training period. She is involved in multiple projects in college which are related to Deep Learning and Machine Learning.
References (32)
- et al.
Improved diagnosis of parkinson’s disease based on optimized crow search algorithm
Comput. Electr. Eng. J.
(2018) - et al.
Optimized cuttlefish algorithm for the diagnosis of Parkinson’s disease
Cognitive Syst. Res. J.
(2018) - et al.
Grasshopper optimisation algorithm: Theory and application
Adv. Eng. Softw.
(2017) Autistic spectrum disorders
BMJ Clin. Res. ed.
(1996)- et al.
Vocal identity recognition in autism spectrum disorder
PLoS One
(2015) - et al.
Sex differences in autism spectrum disorders in current opinion
Neurology
(2013) - et al.
Prevalence of autism spectrum disorder among children aged 8 years — Autism and developmental disabilities monitoring network, 11 sites, United States, 2014
MMWR Surveill. Summ.
(2018) - et al.
Different neurodevelopmental symptoms have a common genetic etiology
J. Child Psychol. Psychiatry
(2013) - et al.
Diagnosis in autism: A survey of over 1200 patients in the UK
Autism
(1997) Machine learning in autistic spectrum disorder behavioral research: A review and ways forward
Inform. Health Soc. Care
(2018)
Use of machine learning for the behavioral distinction of autism and ADHD
Transl. Psychiatry
Efficient autism spectrum disorder prediction with eye movement: A machine learning framework
Searching for a minimal set of behaviors for autism detection through feature selection-based machine learning
Transl. Psychiatry
Use of machine learning to shorten observation-based screening and diagnosis of autism
Transl. Psychiatry
Hybrid grasshopper optimization algorithm and support vector machines for automatic seizure detection in EEG signals
Adv. Intell. Syst. Comput.
Chronological grasshopper optimization algorithm-based gene selection and Cancer classification
J. Adv. Res. Dynam. Control Syst.
Cited by (28)
Autism spectrum disorder detection and classification using chaotic optimization based Bi-GRU network: An weighted average ensemble model
2023, Expert Systems with ApplicationsTowards finding the lost generation of autistic adults: A deep and multi-view learning approach on social media
2023, Knowledge-Based SystemsFundamentals of machine-learning modeling for behavioral screening and diagnosis of autism spectrum disorder
2022, Neural Engineering Techniques for Autism Spectrum Disorder: Volume 2: Diagnosis and Clinical AnalysisAdvancing ASD detection: novel approach integrating attention graph neural networks and crossover boosted meerkat optimization
2024, International Journal of Machine Learning and CyberneticsAn ensemble classification and binomial cumulative based PCA for diagnosis of parkinson’s disease and autism spectrum disorder
2024, International Journal of System Assurance Engineering and Management
Miss. Nikita Goel was born in India (1997) and currently residing in Delhi. She is pursuing her B.Tech. in Computer Science and Engineering from Maharaja Agrasen Institute of Technology (2016–2020), GGSIPU, Delhi. She has been conferred on as the Senior Under Officer in NCC. She has gained expertise in Python, Machine Learning and is improving her Deep Learning skills in her training period. She is involved in multiple projects in college which are related to Deep Learning and Machine Learning. Her research area includes Machine Learning and Deep Learning especially in the medical domain.
Miss. Bhavya Grover was born in India (1998) and currently residing in Delhi. She is pursuing her B.Tech. in Computer Science and Engineering from Maharaja Agrasen Institute of Technology (2016-2020), GGSIPU, Delhi. She is learning Python, Machine Learning and is improving her Java skills in her training period. She is involved in multiple projects in college which are related to Machine Learning. Her research area includes Machine Learning and Deep Learning. She was also an active member of a non-profit organization named Leaders for Tomorrow where she helped them organize various awareness programs.
Mr. Anuj was born in India (1997). He is pursuing her B.Tech. in Computer Science and Engineering from Maharaja Agrasen Institute of Technology (2016-2020), GGSIPU, Delhi. He has gained expertise in Java, Android Development and is improving his Machine Learning and Deep Learning skill in his training period. He is involved in multiple projects in college which are related to Machine Learning. His research are includes Machine Learning and Deep Learning. He was also an active member of IOSD (International Organization of Software Developers).
Dr. Deepak Gupta is an eminent academician; plays versatile roles and responsibilities juggling between lectures, research, publications, consultancy, and community service. With 12 years of rich expertise in teaching and two years in the industry; he focuses on rational and practical learning. He has served as Editor-in-Chief, Guest Editor, Associate Editor in SCI and various other reputed journals. He has actively been an organizing end of various reputed International conferences. He has completed his Post-Doc from Inatel, Brazil, and Ph.D. from Dr. APJ Abdul Kalam Technical University. He has authored/Edited 37 books with National/International level publisher. He has published 92 research publications in reputed International Journals and Conferences including 45 SCI Indexed Journals. Invited as a Faculty Resource Person/Session Chair/Reviewer/TPC member in different FDP, conferences, and journals.
Dr. Ashish Khanna has expertise in Teaching, Entrepreneurship, and Research & Development. He received his Ph.D. degree from National Institute of Technology, Kurukshetra in March 2017. He has completed his M. Tech. and B. Tech. from GGSIPU, Delhi. He has completed his PDF from Internet of Things Lab at Inatel, Brazil. He has around 90 accepted and published research works in reputed SCI, Scopus journals, conferences and reputed book series including around 40 papers in SCI indexed Journals with cumulative impact factor of above 100. Additionally, He has authored, edited and editing 20 books. He is Series Editor in De Gruyter (Germany) of ”Intelligent Biomedical Data Analysis” series. His research interest includes image processing, Distributed Systems and its variants (MANET, FANET, VANET, IoT), Machine learning, Evolutionary computing and many more.
Mr. Moolchand Sharma is currently an Assistant Professor in the Department of Computer Science and Engineering at Maharaja Agrasen Institute of Technology, GGSIPU Delhi. He has published scientific research publications in reputed International Journals and Conferences including SCI indexed and Scopus indexed Journals such as Cognitive Systems Research (Elsevier), International Journal of Image & Graphics (World Scientific),International Journal of Innovative Computing and Applications (Inderscience) & Innovative Computing and Communication Journal (Scientific Peer-reviewed Journal). He has authored/co-authored in chapters with International publishers like Elsevier, Wiley, De Gruyter. His research area includes Artificial Intelligence, Nature-Inspired Computing, Security in Cloud Computing, Machine Learning and Search Engine Optimization. He is associated with various professional bodies like IAENG, ICSES, UACEE, Internet Society etc. He possesses teaching experience of more than 7 years. He is also the co-convener of ‘ICICC’ springer conference series. He is a doctoral researcher at DCR University of Science & Technology, Haryana. He completed his Post Graduate in 2012 from SRM UNIVERSITY, NCR CAMPUS, GHAZIABAD, and Graduate in 2010 from KNGD MODI ENGG. COLLEGE, GBTU.