Unlocking the Potential of Autism Detection: Integrating Traditional Feature Selection and Machine Learning Techniques DOI Open Access

Samar Hazim Hammed,

A. S. Albahri

Applied Data Science and Analysis, Journal Year: 2023, Volume and Issue: unknown, P. 42 - 58

Published: May 1, 2023

The diagnostic process for Autism Spectrum Disorder (ASD) typically involves time-consuming assessments conducted by specialized physicians. To improve the efficiency of ASD screening, intelligent solutions based on machine learning have been proposed in literature. However, many existing ML models lack incorporation medical tests and demographic features, which could potentially enhance their detection capabilities considering affected features through traditional feature selection approaches. This study aims to address aforementioned limitation utilizing a real dataset containing 45 983 patients. achieve this goal, two-phase methodology is employed. first phase data preparation, including handling missing model-based imputation, normalizing using Min-Max method, selecting relevant approaches features. In second phase, seven classification techniques recommended literature, Decision Trees (DT), Random Forest (RF), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), AdaBoost, Gradient Boosting (GB), Neural Network (NN), are utilized develop models. These then trained tested prepared evaluate performance detecting ASD. assessed various metrics, such as Accuracy, Recall, Precision, F1-score, AUC, Train time, Test time. metrics provide insights into models' overall accuracy, sensitivity, specificity, trade-off between true positive false rates. results highlight effectiveness Specifically, GB model outperforms other with an accuracy 87%, Recall Precision 86%, F1-score AUC 95%, time 21.890, 0.173. Additionally, benchmarking analysis against five studies reveals that achieves perfect score across three key areas. By approaches, developed demonstrate improved potential screening diagnosis processes.

Language: Английский

An evaluation of machine learning approaches for early diagnosis of autism spectrum disorder DOI Creative Commons

Rownak Ara Rasul,

Promy Saha,

Diponkor Bala

et al.

Healthcare Analytics, Journal Year: 2024, Volume and Issue: 5, P. 100293 - 100293

Published: Jan. 4, 2024

Autistic Spectrum Disorder (ASD) is a neurological disease characterized by difficulties with social interaction, communication, and repetitive activities. While its primary origin lies in genetics, early detection crucial, leveraging machine learning offers promising avenue for faster more cost-effective diagnosis. This study employs diverse methods to identify crucial ASD traits, aiming enhance automate the diagnostic process. We eight state-of-the-art classification models determine their effectiveness detection. evaluate using accuracy, precision, recall, specificity, F1-score, area under curve (AUC), kappa, log loss metrics find best classifier these binary datasets. Among all models, children dataset, SVM LR achieve highest accuracy of 100% adult model produces 97.14%. Our proposed ANN provides 94.24% new combined dataset when hyperparameters are precisely tuned each model. As almost high which utilize true labels, we become interested delving into five popular clustering algorithms understand behavior scenarios without labels. calculate Normalized Mutual Information (NMI), Adjusted Rand Index (ARI), Silhouette Coefficient (SC) select models. evaluation finds that spectral outperforms other benchmarking terms NMI ARI while demonstrating comparability optimal SC achieved k-means. The implemented code available at GitHub.

Language: Английский

Citations

19
A Systematic Review of Using Deep Learning Technology in the Steady-State Visually Evoked Potential-Based Brain-Computer Interface Applications: Current Trends and Future Trust Methodology DOI Creative Commons
A. S. Albahri, Z.T. Al-Qaysi, Laith Alzubaidi

et al.

International Journal of Telemedicine and Applications, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 24

Published: April 30, 2023

The significance of deep learning techniques in relation to steady-state visually evoked potential- (SSVEP-) based brain-computer interface (BCI) applications is assessed through a systematic review. Three reliable databases, PubMed, ScienceDirect, and IEEE, were considered gather relevant scientific theoretical articles. Initially, 125 papers found between 2010 2021 related this integrated research field. After the filtering process, only 30 articles identified classified into five categories on their type methods. first category, convolutional neural network (CNN), accounts for 70% (n = 21/30). second recurrent (RNN), 10% 3/30). third fourth categories, (DNN) long short-term memory (LSTM), account 6% 30). fifth restricted Boltzmann machine (RBM), 3% 1/30). literature's findings terms main aspects existing pattern recognition SSVEP-based BCI, such as feature extraction, classification, activation functions, validation methods, achieved classification accuracies, are examined. A comprehensive mapping analysis was also conducted, which six categories. Current challenges ensuring trustworthy BCI discussed, recommendations provided researchers developers. study critically reviews current unsolved issues development selection multicriteria decision-making (MCDM). trust proposal solution presented with three methodology phases evaluating benchmarking using fuzzy techniques. Valuable insights developers provided.

Language: Английский

Citations

28
Role of Artificial Intelligence for Autism Diagnosis Using DTI and fMRI: A Survey DOI Creative Commons
Eman Mohamed Helmy, Ahmed Elnakib, Yaser ElNakieb

et al.

Biomedicines, Journal Year: 2023, Volume and Issue: 11(7), P. 1858 - 1858

Published: June 29, 2023

Autism spectrum disorder (ASD) is a wide range of diseases characterized by difficulties with social skills, repetitive activities, speech, and nonverbal communication. The Centers for Disease Control (CDC) estimates that 1 in 44 American children currently suffer from ASD. current gold standard ASD diagnosis based on behavior observational tests clinicians, which being subjective time-consuming afford only late detection (a child must have mental age at least two to apply an observation report). Alternatively, brain imaging-more specifically, magnetic resonance imaging (MRI)-has proven its ability assist fast, objective, early detection. With the recent advances artificial intelligence (AI) machine learning (ML) techniques, sufficient tools been developed both automated More recently, development deep (DL), young subfield AI neural networks (ANNs), has successfully enabled processing MRI data improved diagnostic abilities. This survey focuses role autism diagnostics basic modalities: diffusion tensor (DTI) functional (fMRI). In addition, outlines findings DTI fMRI autism. Furthermore, techniques using are summarized discussed. Finally, emerging tendencies described. results this study show how useful early, diagnosis. solutions potential be used healthcare settings will introduced future.

Language: Английский

Citations

26
Latest clinical frontiers related to autism diagnostic strategies DOI Creative Commons
Samuele Cortese, Alessio Bellato, Alessandra Gabellone

et al.

Cell Reports Medicine, Journal Year: 2025, Volume and Issue: unknown, P. 101916 - 101916

Published: Jan. 1, 2025

Language: Английский

Citations

1
Diagnosis-Based Hybridization of Multimedical Tests and Sociodemographic Characteristics of Autism Spectrum Disorder Using Artificial Intelligence and Machine Learning Techniques: A Systematic Review DOI Creative Commons
M. E. Alqaysi, A. S. Albahri, Rula A. Hamid

et al.

International Journal of Telemedicine and Applications, Journal Year: 2022, Volume and Issue: 2022, P. 1 - 26

Published: July 1, 2022

Autism spectrum disorder (ASD) is a complex neurobehavioral condition that begins in childhood and continues throughout life, affecting communication verbal behavioral skills. It challenging to discover autism the early stages of which prompted researchers intensify efforts reach best solutions treat this challenge by introducing artificial intelligence (AI) techniques machine learning (ML) algorithms, played an essential role greatly assisting medical healthcare staff trying obtain highest predictive results for disorder. This study aimed at systematically reviewing literature related criteria, including multimedical tests sociodemographic characteristics AI ML contributions. Accordingly, checked Web Science (WoS), Direct (SD), IEEE Xplore digital library, Scopus databases. A set 944 articles from 2017 2021 collected reveal clear picture better understand all academic through definitive collection 40 based on our inclusion exclusion criteria. The selected were divided similarity, objective, aim evidence across studies. They are into two main categories: first category "diagnosis ASD questionnaires features" (

Language: Английский

Citations

33
Artificial intelligence-based approaches for improving the diagnosis, triage, and prioritization of autism spectrum disorder: a systematic review of current trends and open issues DOI

Shahad Sabbar Joudar,

A. S. Albahri, Rula A. Hamid

et al.

Artificial Intelligence Review, Journal Year: 2023, Volume and Issue: 56(S1), P. 53 - 117

Published: June 21, 2023

Language: Английский

Citations

21
Fuzzy Decision-Making Framework for Sensitively Prioritizing Autism Patients with Moderate Emergency Level DOI Open Access

H Talib,

A. S. Albahri, Thierry Oscar Edoh

et al.

Applied Data Science and Analysis, Journal Year: 2023, Volume and Issue: unknown, P. 16 - 41

Published: March 15, 2023

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder that requires careful assessment and management. The prioritization of ASD patients involves navigating through complexities such as conflicts, trade-offs, the importance different criteria. Therefore, this study focuses on prioritizing with in healthcare setting an evaluation benchmarking framework. aim to develop framework utilizes Multi-Criteria Decision Making (MCDM) methods assist professionals patients, particularly those moderate injury levels. methodology outlines several phases, including dataset identification, development decision matrix, weighting 19 criteria using FWZIC method, ranking 432 VIKOR evaluating proposed four sensitivity analysis scenarios. Among criteria, criterion 'verbal communication' obtained highest weight. Additionally, 'laughing for no reason', 'nodding', 'patient movement at home', 'pointing index finger' similar higher weights, indicating their potential impact patients. experimental results highlight significance adjusting weights influencing final rankings method. This emphasizes need consideration when assigning ensure accurate prioritization. Moreover, provides valuable insights into improving care support provided individuals autism Iraq. findings contribute existing body knowledge field pave way future research interventions aimed enhancing quality

Language: Английский

Citations

21
Evaluation of Autism Spectrum Disorder Based on the Healthcare by Using Artificial Intelligence Strategies DOI Creative Commons
Amit Sundas, Sumit Badotra, Shalli Rani

et al.

Journal of Sensors, Journal Year: 2023, Volume and Issue: 2023(1)

Published: Jan. 1, 2023

The behaviors of children with autism spectrum disorder (ASD) are often erratic and difficult to predict. Most the time, they unable communicate effectively in their own language. Instead, using hand gestures pointing phrases. Because this, it can be for caregivers grasp patients’ requirements, although early detection condition make this much simpler. Assistive technology Internet Things (IoT) alleviate absence verbal nonverbal communication community. IoT‐based solutions use machine Learning (ML) deep learning (DL) algorithms diagnose enhance lives patients. A thorough review ASD techniques setting IoT devices is presented research. Identifying important trends health care research primary objective review. There also a technical taxonomy organizing current articles on methodologies based different factors such as AI, SS network, ML, IoT. On basis criteria accuracy sensitivity, statistical operational analyses examined presented.

Language: Английский

Citations

20
Breaking Barriers—The Intersection of AI and Assistive Technology in Autism Care: A Narrative Review DOI Open Access

Antonio Iannone,

Daniele Giansanti

Journal of Personalized Medicine, Journal Year: 2023, Volume and Issue: 14(1), P. 41 - 41

Published: Dec. 28, 2023

(Background) Autism increasingly requires a multidisciplinary approach that can effectively harmonize the realms of diagnosis and therapy, tailoring both to individual. Assistive technologies (ATs) play an important role in this context hold significant potential when integrated with artificial intelligence (AI). (Objective) The objective study is analyze state integration AI ATs autism through review. (Methods) A review was conducted on PubMed Scopus, applying standard checklist qualification process. outcome reported 22 studies, including 7 reviews. (Key Content Findings) results reveal early yet promising interest integrating into assistive technologies. Exciting developments are currently underway at intersection robotics, as well creation wearable automated devices like smart glasses. These innovations offer substantial for enhancing communication, interaction, social engagement individuals autism. Presently, researchers prioritizing innovation over establishing solid presence within healthcare domain, where issues such regulation acceptance demand increased attention. (Conclusions) As field continues evolve, it becomes clear will pivotal bridging various domains, positioned act crucial connectors.

Language: Английский

Citations

20
Intelligent triage method for early diagnosis autism spectrum disorder (ASD) based on integrated fuzzy multi-criteria decision-making methods DOI Creative Commons

Shahad Sabbar Joudar,

A. S. Albahri, Rula A. Hamid

et al.

Informatics in Medicine Unlocked, Journal Year: 2022, Volume and Issue: 36, P. 101131 - 101131

Published: Nov. 16, 2022

Autism spectrum disorder (ASD) symptoms and severity levels vary from patient to patient, so treatment healthcare will vary. However, little attention has been given developing an autistic triage method for ASD patients concerning four issues: hybrid criteria, multi-selection criteria problems, importance, trade-off based on the inverse relationship between criteria. Therefore, this study aims develop a new triaging classifying them according their of using Fuzzy Multi-Criteria Decision Making (fMCDM) methods. Two methodology phases have conducted: first phase is identify preprocess dataset, including 988 with 42 medical Sociodemographic In second phase, two fMCDM methods were used method. The fuzzy Delphi Method (FDM) select most influential among thirteen psychologists in psychological field. Then Fuzzy-Weighted Zero-Inconsistency (FWZIC) assign weights important psychologists' opinions. Accordingly, Processes Triaging Patients (PTAP) developed time into three levels: minor, moderate, urgent. For preprocessed 538 out obtained as dataset underwent data cleaning capture only autism patients. FDM results selected 19 can control bias opinions, FWZIC assigned appropriate PTAP triages 36 minor injuries, 432 moderate 70 urgent injuries. More complex statistical analyses presented MedCalc software. Three physicians field gave subjective judgements diagnosis 46 random samples sensitivity 86.67%, 80%, 90.91%, while specificity 93.55%, 88.46%, 94.29% urgent, levels, respectively. addition, accuracy was 91.30% 84.78% 93.48% minor. This assessment led deduction that proposed be applied high performance. early application support clinical utilizing advantages techniques multidimensional Four psychologists, acquired 15 correlation analysis 'Wave' criterion highest level 0.4523. On contrary, "Pointing index finger" lowest −0.0542. Limitations future works also reported study. confirms efficacy compared previous studies five comparative points 100%.

Language: Английский

Citations

24