Neural Networks Unveiling Autism: Deep Learning Approaches for Spectrum Detection

The neurological condition known as an autism spectrum disorder (ASD) commonly coexists with sensory problems, such as an excessively or insufficiently sensitive reaction to sounds, smells, or touch. Even if heredity is the primary cause, situations may be improved with early discovery and treatment. The study offers a novel method for using a deep classifier to identify ASD. The following stages describe how the identification of ASD works. The screening technique is significantly more effective by employing Random Forest feature analysis to explain ASD symptoms. Moreover, ASD class type and assessment parameters are reported using Machine Learning (ML) classifier models. Combining Random Forest with RNN can enhance the predictive model’s performance for autism spectrum disorder by revealing which features are most important for making predictions, enhancing the interpretability of the model, and boosting its overall performance and generalizability. The accuracy for the whole project is 99%. This suggests that the classifier can correctly predict the autism status of 99% of the family members in the test set.