Jun 2026· Journal of Soft Computing and Data Mining· Vol 7· 0 citations
TL;DR
The findings verify that incorporating deep and traditional learning models in an ensemble framework can significantly enhance the classification of ASD, offering a dependable and scalable computational model to aid clinical specialists in the initial diagnosis of ASD.
Abstract
Conventional diagnostic tests for Autism Spectrum Disorder (ASD) involve the use of subjective behavioral observations and questionnaires completed by the clinician, which can be time-consuming and subjectto human bias. The challenge encourages the development of innovative, data-driven methods to facilitateearly and accurate identification of ASD. The research proposes a Multi-Model Ensemble Approach Using Deep and traditional learning for ASD classification (MME-ASD) model. The MME-ASD model encompassesthree traditional machine learning (ML) and two deep learning (DL) algorithms that perform according to a weighted majority voting strategy. The five learning paradigms are Random Forest (RF), Decision Tree (DT), Neural Networks(NN), Convolutional Neural Networks(CNN), and Deep Recurrent Neural Networks(DRNN),which are utilized to enhance classification accuracy and generalization. An ensemble evaluation method is proposed to complete this study andassess the efficiency of the proposed MME-ASD model. The MME-ASD model acquires complementary properties by using numeric and textual data from a publicly available dataset of ASD, which includes information on 704 adults, both with and without a diagnosis. Initially, during the evaluation phase, the performance of the standalone traditional ML and DL algorithms was assessed acrossseveral train-test ratios. Subsequently, the proposed MME-ASD ensemble was evaluated with a 60-40 split to ensure compatibility with the baseline models. Finally, a 3-fold cross-validation experiment was conducted to assess the robustness and generalization of the proposed MME-ASD model. The experimental outcomes reveal that the MME-ASD model outperformsindividual learners for both cross-validation and train-test assessments. It records evaluation metrics of accuracy 99.57%,precision 99.48%, and recall 98.94% across the 3-fold cross-validation experiments. The findings verify that incorporating deep and traditional learning models in an ensemble framework can significantly enhance the classification of ASD, offering a dependable and scalable computationalmodel to aid clinical specialists in the initial diagnosis of ASD.
The need to develop large, well‐balanced datasets, the application of explainable AI techniques, standardization and regulatory guidelines for facilitating the clinical translation of ASD detection systems are suggested.
Anupama N, Chandrashekar M. Patil· International Journal of Dev...· 0 citations
Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental condition characterized by diverse behavioral, cognitive, sensory, and communication profiles, making early diagnosis and personalized intervention challenging. Recent advances in machine learning (ML) and deep learning (DL) have enabled the development of computational tools for ASD screening, classification, severity assessment, and intervention monitoring. This review synthesizes findings from 50 recent studies that applied ML and DL techniques to ASD-related datasets, including electroencephalography (EEG), eye-tracking, behavioral video, microbiome, voice acoustic, demographic, and multimodal data. The review addresses three key questions: (i) which data modalities and computational approaches are most frequently used, (ii) how diagnostic performance is evaluated across different study designs, and (iii) what methodological challenges limit clinical translation. The literature is organized according to data modality, algorithmic approach, and clinical readiness. Approaches examined include conventional ML methods, convolutional neural networks, graph neural networks, hybrid deep learning architectures, federated learning, explainable artificial intelligence, topological data analysis, and multimodal fusion. The findings suggest that multimodal and graph-based approaches provide a more comprehensive representation of ASD phenotypes than single-modality methods. Explainability and privacy-preserving learning have also emerged as important considerations for clinical deployment. However, many reported high-performance models are based on small sample sizes, repeated use of the ABIDE dataset, class imbalance, single-site validation, or limited external testing, raising concerns regarding generalizability. Beyond diagnostic accuracy, this review evaluates model interpretability, calibration, scalability, validation rigor, and clinical applicability. Overall, the analysis highlights the need for standardized benchmarks, externally validated multimodal datasets, clinically relevant evaluation metrics, and decision-support systems that complement rather than replace expert clinical assessment in ASD diagnosis and management.
S. K, Lakshmi Annapurna Y· Journal of Visualized Experi...· 0 citations
Autism spectrum disorder (ASD) is a complex neurodevelopmental condition characterized by persistent deficits in social communication, interaction, and behavioral patterns. Early screening is essential for timely intervention; however, conventional diagnostic procedures are often resource-intensive, subjective, and dependent on clinical expertise. This study proposes an integrated hybrid artificial intelligence framework for preliminary ASD screening by combining machine learning and deep learning techniques using structured behavioral data and facial image analysis. The first module employs a Random Forest classifier to analyze Autism Quotient-10 (AQ-10) questionnaire responses together with demographic and medical history attributes, including age, gender, ethnicity, family history, and developmental factors. The second module utilizes ResNet18 with transfer learning to classify facial images after preprocessing and data augmentation, enabling non-invasive image-based screening. Both predictive models are integrated into a user-friendly screening application developed using CustomTkinter, allowing independent questionnaire-based and image-based assessments through a unified graphical interface. Model training incorporates feature preprocessing, hyperparameter optimization using Grid Search with cross-validation, and performance evaluation using accuracy, precision, recall, F1-score, specificity, and area under the receiver operating characteristic curve (ROC-AUC). Experimental results demonstrate that the Random Forest model achieved an accuracy of 96.8%, while the ResNet18 model attained 94.2% accuracy, indicating the effectiveness of combining behavioral and facial information for preliminary ASD screening. The proposed framework is intended as a clinical decision-support tool to facilitate accessible and timely screening and is not designed to replace comprehensive clinical diagnosis.
Neha A. Kandalkar, R. Jogekar· Adolescência e Saúde· 0 citations
Early detection of autism spectrum disorder (ASD) is essential for timely intervention. This study presents a hybrid artificial intelligence framework for non-invasive ASD pre-screening using children’s coloring, drawing, and handwriting activities. The proposed framework combines deep convolutional neural networks (VGG16, ResNet50, and EfficientNetB0) as feature extractors with a support vector machine (SVM) classifier to distinguish four diagnostic categories: non-ASD, mild ASD, moderate ASD, and severe ASD. Experimental results demonstrate task-specific performance across architectures. ResNet50–SVM achieved perfect classification for coloring tasks, with 100% accuracy, precision, recall, and F1-score. VGG16–SVM performed best for drawing, achieving 88% accuracy and recall, 89% precision, and an F1-score of 87%. EfficientNetB0–SVM produced the highest handwriting performance, achieving 96% across all evaluation metrics. These findings demonstrate the potential of computer vision-based analysis of children’s expressive activities as an effective, non-invasive ASD pre-screening tool. Future work will focus on expanding dataset diversity and integrating multimodal behavioral cues to improve model generalization and clinical applicability.
Aina Khairina Ahmad Khair, Wan Mohd Yaakob Wan Bejuri, Mohd Murtadha Mohamad et al.· Bulletin of Electrical Engin...· 0 citations
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder which necessitates early and precise diagnosis to facilitate prompt intervention and enhance long-term results. Using behavioural and demographic data, this research offers a machine learningbased method for identifying ASD. The proposed system incorporates comprehensive data preprocessing techniques, such as addressing missing values, encoding categorical features, and feature scaling, to increase model efficiency. Additionally, the most pertinent characteristics that contribute to the classification of ASD are found using feature selection techniques. The predictive power of four machine learning (ML) algorithms: Support Vector Machine (SVM), Random Forest (RF), Cat-Boost (CB), and Light Gradient Boosting Machine (LGBM) is assessed through implementation. Following training and validation using appropriate data partitioning methodologies, the model's performance is evaluated using different performance metrics. A comparative study is undertaken to find out how well every model represents complex trends in the dataset. According to the experimental results, boosting-based methods in particular, Cat-Boost and LGBM perform better because they can effectively handle categorical variables and minimise overfitting. Cat-Boost achieved best results with an accuracy of 85.14%. This study shows how cutting-edge machine learning methods can improve ASD screening and assist with data-driven clinical decision-making.
Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition characterized by significant heterogeneity, making the accurate classification of its severity levels crucial for effective intervention. This study develops and validates a comprehensive machine learning framework to classify ASD severity (mild, moderate, severe) by investigating the differential impact of feature engineering and selection. A dataset of 340 individuals was analyzed using a dual-framework approach, integrating supervised classification (SVM, Random Forest, XGBoost) and unsupervised clustering (K-Means, Hierarchical, GMM). The methodology centered on comparing a feature-centric approach, using a hybrid selection strategy on 60 engineered and original features, against a baseline approach using only 33 original features. The feature-centric approach yielded markedly superior results; a Random Forest classifier, trained on a minimal subset of just nine engineered features, achieved a test accuracy of 76.5%, significantly outperforming all other models, including an SVM trained on 22 original features (70.6% accuracy). This highlights that feature quality is more critical than quantity. The unsupervised analysis revealed a critical “evaluation paradox,” where radical, unguided feature reduction improved geometric cluster cohesion but degraded clinical accuracy. Conversely, a guided, domain-informed selection improved both internal and external metrics.
Arazo Mohammed Mustafa· ARID International Journal f...· 0 citations