<![CDATA[Diagnosing Autism Spectrum Disorder (ASD) in adults is a challenging task, requiring precise and efficient early detection methods. However, there is limited research in this area. Hence, this study seeks to address this gap by evaluating the effectiveness of Convolutional Neural Networks (CNNs) compared to traditional machine learning techniques for detecting autism in adults. The study introduces a CNN-based model and conducts a performance comparison with conventional algorithms such as Support Vector Machine (SVM), Decision Tree, K-Nearest Neighbors (KNN), Gaussian Naive Bayes (GNB), and Gradient Boosting (GBoost). The objectives are to evaluate the efficacy of CNNs in adult autism detection, identify algorithm strengths and weaknesses, and explore healthcare implications. The research utilizes the Autism Screening on Adults dataset, with 704 records and 21 features, employing preprocessing steps to optimize data quality. The proposed CNN model encompasses convolutional layers, max-pooling, dropout, and dense layers, while baseline algorithms serve as benchmarks. Evaluation metrics include the Confusion Matrix and Classification Report. The CNN model achieved remarkable accuracy (99%) and precision in adult autism detection, outperforming traditional algorithms. SVM emerged as the closest competitor but fell short. This study underscores CNN's potential for precise autism detection in adults, with implications for early intervention and telehealth applications. The research highlights CNNs' effectiveness and superiority over traditional machine learning algorithms, suggesting their promise for accurate diagnosis. Future research opportunities include expanding datasets, optimizing model parameters, and addressing ethical considerations for practical healthcare implementation]]>
