Explainable AI based ensemble model for the identification of Schizophrenia prodromal phase

2026-04-18 · Source: Machine learning : nature.com subject feeds · Field: Health & Wellbeing — Mental Health & Psychological Support, Medical Devices & Health Technology, Health & Medical Research · Depth: Advanced, short

Summary

Researchers developed an AI framework utilizing Machine Learning (ML) and Ensemble Learning (EL) models with Feature Selection to predict prodromal symptoms in Schizophrenia patients. This framework processed an open-source dataset from 5000 patients, encompassing clinical, psychological, and behavioral symptoms. The customized EL-based STACK models achieved high performance metrics, including 96.2% Accuracy, 96% Precision, 96% Recall, 96.7% F1-score, and 93% Average AUC. To enhance interpretability, the study integrated Explainable AI (XAI) techniques, specifically the Shapley Additive exPlanations (SHAP) architecture. SHAP generated visualizations like Violin, Waterfall, Force, and Dependence Plots, providing meaningful interpretations of the classifier's predictions to assist clinicians in informed decision-making regarding early Schizophrenia detection.

Key takeaway

For clinicians and AI scientists developing diagnostic tools, this research demonstrates that integrating Explainable AI with ensemble machine learning can significantly improve the accuracy and interpretability of early Schizophrenia detection. You should consider adopting SHAP-based visualizations to provide transparent insights into model predictions, thereby fostering trust and facilitating more informed clinical decisions regarding prodromal symptom identification.

Key insights

An XAI-integrated ensemble ML model accurately predicts Schizophrenia prodromal symptoms for early clinical intervention.

Principles

• Ensemble learning improves predictive performance.
• XAI enhances model interpretability for clinical use.

Method

The method involves using ML and EL models with feature selection on clinical, psychological, and behavioral data, followed by SHAP for explainability via various visualizations.

In practice

• Apply SHAP for model interpretability.
• Utilize ensemble stacking for high accuracy.

Topics

• Schizophrenia Prodromal Phase
• Explainable AI
• Ensemble Learning
• Machine Learning Models
• SHAP

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• Explainable AI in Production: A Neuro-Symbolic Model for Real-Time Fraud Detection — Integrating explanations directly into model architecture enables real-time, deterministic, and low-latency explainability.
• Application of integrated gradients explainability to sociopsychological semantic markers — Integrated gradients reveal word-level contributions to sociopsychological text classification, even with limited data.
• Interpretable and Explainable Surrogate Modeling for Simulations: A State-of-the-Art Survey and Perspectives on Explainable AI for Decision-Making — Integrating XAI with surrogate models enhances transparency and diagnostic capabilities for complex simulations, bridging efficiency and interpretability.
• Towards Rigorous Explainability by Feature Attribution — Existing SHAP scores for ML explainability are flawed and misleading; rigorous alternatives are necessary.
• Explainable Fall Detection for Elderly Care via Temporally Stable SHAP in Skeleton-Based Human Activity Recognition — Temporally stable SHAP attributions improve reliability for explainable fall detection in elderly care.
• Translating black-box medical AI models into interpretable global decision logic — Class-Association Manifold Learning (CAML) explains black-box medical AI by visualizing global decision rules on low-dimensional manifolds.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.