Authors: Trapti Shrivastava, Harshal Chaudhari, Vrijendra Singh

Published on: April 02, 2024

Impact Score: 7.6

Arxiv code: Arxiv:2404.02181

Summary

• What is new: The development of a simpler, quicker, and cheaper ML-based technique for identifying ASD utilizing a minimized set of questions with higher accuracy than existing methods.
• Why this is important: Existing clinical screening tests for ASD are expensive and time-consuming, and previous ML techniques haven’t effectively predicted autistic features using the Indian ASD database.
• What the research proposes: A machine learning model utilizing various classifiers to predict ASD from minimally required input, tested on the AIIMS Modified INDT-ASD database.
• Results: The SVM model achieved the highest accuracy (100%) in predicting ASD, outperforming other models in accuracy and recall.

Technical Details

Technological frameworks used: nan

Models used: Adaboost, Gradient Boost, Decision Tree, Logistic Regression, Random Forest, Gaussian Naive Bayes, Linear Discriminant Analysis, Quadratic Discriminant Analysis, K-Nearest Neighbors, Support Vector Machine

Data used: AIIMS Modified INDT-ASD database

Potential Impact

Healthcare providers, ASD diagnosis centers, educational technology companies focusing on special education

Want to implement this idea in a business?

We have generated a startup concept here: ASDEarlyScan.