Python
Aim:
The aim of this work is to develop an accurate and interpretable machine learning framework for early-stage detection of Autism Spectrum Disorder (ASD) by integrating explainable artificial intelligence techniques to enhance clinical trust and decision transparency.
Abstract:
Early identification of Autism Spectrum Disorder (ASD) is essential for effective intervention and improved developmental outcomes. This work presents an advanced machine learning–based framework for early-stage ASD detection using structured behavioral screening data. The proposed system employs comprehensive data preprocessing, feature encoding, class balancing, and optimized model training to achieve reliable prediction performance. To overcome the black-box nature of conventional machine learning models, Explainable Artificial Intelligence (XAI) is incorporated using SHapley Additive exPlanations (SHAP). The explainability module provides both global and instance-level interpretations by identifying key behavioral features influencing model decisions. Experimental results demonstrate improved predictive accuracy along with enhanced interpretability, making the system suitable for real-world clinical decision support. The integration of SHAP significantly improves transparency and supports clinicians in understanding and validating model predictions.
Proposed System:
The proposed system introduces an advanced machine learning framework combined with Explainable AI. After preprocessing and model optimization, SHAP is applied to explain individual predictions and global feature importance. This approach ensures both high predictive performance and transparent decision-making.
Advantage:
- Improved prediction accuracy
- Transparent and interpretable model decisions
- Feature-level explanation using SHAP
- Increased clinical trust and usability
- Suitable for early-stage ASD screening
Reviews
There are no reviews yet.