BACKGROUND OF INVENTION:
In recent years, the incidence of Autism Spectrum Disorder (ASD) has seen a noticeable rise, prompting global concern regarding early detection and intervention. ASD is a developmental disorder that affects a person's ability to communicate, interact socially, and exhibit typical behavioral patterns. Studies have consistently shown that early diagnosis and timely therapy significantly improve the long-term development and quality of life of individuals witli
autism.

Despite the critical importance of early detection, traditional diagnostic processes are often time-consuming, expensive, and dependent on highly specialized healthcare professionals. In many regions—particularly in rural and underserved communities—access to such resources is limited. As a result, there is often a delay in the diagnosis and treatment of affected children, which can severely impact their developmental trajectory.
Technological advancements in artificial intelligence (Al) and machine learning (ML) offer promising alternatives for improving autism screening. By analyzingbehavioral, psychological, and questionnaire-based data, ML models can uncover hidden patterns and predict autism risk with .significant accuracy.
Such tools not only accelerate the diagnostic process but also make it more accessible to non-specialists, such as school counselors, primary care physicians, and parents.
The present invention proposes a novel, interactive, and user-friendly platform that applies machine learning algorithms for the early prediction of autism in children. Built using a combination of Python, Streamlit, and Jupyter Notebook environments, the system allows users to input behavioral traits via questionnaires and receive real-time predictions regarding ASD risk. The platform also provides educational insights, awareness content, and recommendations for further clinical consultation.
This invention aims to bridge the gap in early autism diagnosis by offering an affordable, accessible, and scalable digital screening solution that empowers both professionals and caregivers.
SUMMARY OF INVENTION:
The present invention provides an intelligent software-based system designed for the early prediction of Autism Spectrum Disorder (ASD) using artificial

intelligence and machine learning techniques. The invention leverages behavioral data input through structured questionnaires to identify potential early indicators of autism in children aged between l.5 to 16 years.
The system operates by collecting user responses related to social interaction, communication skills, behavioral patterns, and developmental milestones. These inputs are processed through a trained machine learning model which has been developed using real-world autism screening datasets. The model then classifies the responses to predict the likelihood of the individual being at risk for ASD.
This invention is implemented as a web-based and/or desktop application, providing accessibility for parents, teachers, and healthcare professionals. The system includes a clean user interface for easy data entry and produces a prediction result along with a risk level (e.g., low, moderate, or high) based on the model’s output.
Key features of the invention include:
• Al-powered prediction based on established autism screening criteria • User-friendly form for input collection, designed to minimize complexity • Real-time result generation with high accuracy • Suitable for integration in schools, clinics, and home environments • Enhances awareness and encourages early consultation with specialists By enabling earlier detection, this invention aims to support quicker access to intervention resources, which is crucial for improving developmental outcomes in children diagnosed with autism.

CLAIM
We claim that,

. n intelligent system for the early prediction of Autism Spectrum Disorder (ASD), comprising:
• a user interface for collecting behavioral and developmental data via structured questionnaires; • a trained machine learning model that analyzes the input data; • a prediction module that classifies the data into categories indicating the risk level of ASD.
2. Wherein the machine learning model is selected from supervised learning algorithms such as Decision Trees, Random Forest, Support Vector Machine (SVM), or Neural Networks.
3. Wherein the prediction is based on key behavioral indicators including but not limited to social interaction, repetitive behaviors, communication delays, and sensory sensitivity.
4. Wherein the output comprises a diagnosis label such as “At Risk” or “Not at Risk,” along with a confidence score or probability. 5. Wherein the questionnaire follows standard screening frameworks like M- CHAT or Q-CHAT, adapted for machine-readable input. 6. Wherein the application is implemented as a web application or standalone desktop application compatible with multiple devices. 7. Further comprising a data storage module for maintaining user history, screening records, and prediction outcomes securely. 8.Wherein the software is intended for use by parents, educators, or healthcare professionals for preliminary screening and awareness purposes.

Detailed Description of the Drawings:
The drawing associated with the present invention illustrates the step-by-step flow of the system designed for the early prediction of autism disorder using machine learning techniques. The process begins with the User Interface, where users, such as parents, guardians, or healthcare providers, input essential data through a form-based interface. This input data typically includes answers

to standardized autism screening questionnaires, such as the AQ-10 or M- CHAT, along with relevant demographic details.
Once data is collected, it is passed to the Data Preprocessing Unit, where various operations such as cleaning, encoding categorical variables, normalizing data, and handling missing values are performed to ensure the dataset is consistent and ready for analysis. Following preprocessing, the system moves to Feature Selection, where important features that significantly contribute to autism prediction are identified and retained. This step improves the model's performance and reduces unnecessary complexity.
The Model Training and Selection stage involves feeding the cleaned and selected data into various machine learning algorithms such as Logistic Regression, Support Vector Machines (SVM), or Random Forest classifiers.
These models are trained to recognize patterns and classify input cases as either autistic or non-autistic. Once the models are trained, the Prediction Module utilizes them to make real-time predictions based on new data provided by the
user.