Autism Spectrum Disorders, a neurodevelopmental condition, have surged in prevalence among children. Research suggests that early identification and intervention therapies greatly contribute to favourable long-term outcomes. Detecting any ailment in its initial stages can be challenging, as an initial symptom might indicate various other potential conditions. Top of Form Identifying neurodevelopmental diseases presents an added challenge as their symptoms often evade standard tests, relying instead on behavioural observations for diagnosis. Autism, a prime example of such a disorder, may surface in children as early as one to two years old. Early intervention during this critical period can significantly enhance a patient's condition. The primary focus of the proposed project revolves around utilizing machine learning techniques for detecting and diagnosing Autism, especially when symptoms are incomplete or insufficient, with a specific emphasis on rural areas. Consequently, health assistants or less seasoned doctors often handle such cases. Numerous previously developed systems have utilized complete or partial symptom sets for Autism diagnosis, yet they frequently result in misidentification. The proposed framework employs machine learning techniques to detect and diagnose Autism symptoms in young children, specifically when only incomplete sets of symptoms are accessible