The Toddler Autism Prediction App leverages machine learning to assess the likelihood of Autism Spectrum Disorder (ASD) in toddlers. Designed to be user-friendly for parents and caregivers, the app features an intuitive interface, streamlined data input, and robust privacy protections.

• Overview
• Features
• Prerequisites
• Installation
• Methodology
• Results
• Contributors
• License

Early detection of Autism Spectrum Disorder (ASD) can significantly improve outcomes by providing early interventions and tailored resources for children. This app utilizes a Naive Bayes classifier to predict the likelihood of autism in toddlers based on behavioral and developmental data. The model is designed to be interpretable and accurate, offering caregivers insights into their child’s developmental health.

We began this project by exploring several key research papers related to Graphical Models for Health Diagnosis, including:

• Detection of Autism Spectrum Disorder in Children Using Machine Learning Techniques
• An accessible and efficient autism screening method for behavioural data and predictive analyses
• Employing Bayesian Networks for the Diagnosis and Prognosis of Diseases
• Bayesian Networks for the Diagnosis and Prognosis of Diseases: A Scoping Review
• An Intelligent Bayesian Hybrid Approach to Help Autism Diagnosis* by Paulo Vitor de Campos Souza et al
• Impact of Bayesian Network Model Structure on the Accuracy of Medical Diagnostic Systems

Through this research, we gained foundational insights into Bayesian Networks, understanding their mathematical underpinnings, advantages, challenges, and applications in healthcare.One of the pivotal study, Detection of Autism Spectrum Disorder in Children Using Machine Learning Techniques.

• User-Friendly Interface:
  • Easy navigation and clear prompts for data entry.
• Comprehensive Data Collection:
  • Gathers demographic info, family history, health indicators (e.g., jaundice), and behavioral assessments.
• Behavioral Questionnaire:
  • Includes questions evaluating developmental behaviors like eye contact and social interaction.
• Machine Learning Predictions:
  • Uses a Naive Bayes classifier to analyze user input and predict ASD likelihood.
• Clear Results and Insights:
  • Provides probability-based predictions with interpretative insights.
• Educational Resources:
  • Links to information on autism signs and early detection importance.

• Python 3.x installed on your machine.
• Knowledge of MERN, Flask.
• Basic knowledge of terminal commands.

1. Clone the repository:

   git clone https://github.com/ayushkumar912/Toddler_Autism_Prediction.git
   cd Toddler_Autism_Prediction

2. Create a virtual environment:

   • macOS/Linux

   python3 -m venv .venv

   • Windows

   python -m venv .venv

3. Activate the virtual environment:

   • macOS/Linux

   source .venv/bin/activate 

   • Windows

   .venv\Scripts\activate 

4. Install the required modules:

   pip install -r modules.txt

To run the project, use the following command:

• macOS/Linux

   python3 app.py

• Windows

    python app.py

1. Build the Docker image:

   docker build -t flask-app .

2. Run the container:

      docker run -p 80:80 flask-app

1. Data Collection and Preparation:
   • We used an autism screening dataset from Kaggle. We cleaned and preprocessed the data, addressing missing values and normalizing attributes.
2. Modeling with Naive Bayes:
   • Our initial classifier was Naive Bayes, chosen for its simplicity and effectiveness in handling probabilistic predictions. We also experimented with Random Forest and Ensemble Models for comparison.
3. Frontend Implementation:
   • To showcase the Naive Bayes classifier, we built a basic frontend using HTML, CSS, and Vanilla JS.
4. Testing and Evaluation:
   • We compared model accuracy across Naive Bayes, Random Forest, and Ensemble Models to determine the most reliable predictor.
5. The whole project can be viewed on this repository.

The app provides probabilistic predictions with an emphasis on interpretability, helping caregivers understand which factors significantly influence autism likelihood.

We referred to a wide range of resources throughout the project, including:

• QCHAT-10 Autism Survey for Toddlers
• Kaggle Learning Modules
• Machine Learning Tutorial on YouTube
• ASD Tests Online Resource

• Aninda Paul      Roll No: 202211001
• Ayush Kumar    Roll No: 202211008
• Devrikh Jatav   Roll No: 202211018
• Inarat Hussain  Roll No: 202211030

This project is licensed under the Apache License. See the LICENSE file for details.