Abstract :

The Rising Frequency And Severity Of Dengue Outbreaks Demand Advanced Predictive Systems Capable Of Early Detection, Precise Severity Assessment, And Localized Outbreak Surveillance. Existing Models Predominantly Rely On Static Pathological Thresholds And Conventional Classifiers, Often Lacking Temporal Awareness, Spatial Intelligence, And Clinical Explainability. These Limitations Hinder Their Real-world Deployment In Dynamic Clinical And Public Health Environments. To Address These Gaps, This Study Proposes A Novel Multi-layered Analytical Framework For Dengue Outbreak Detection Using Pathological Metrics, SVM, And XAI, Integrating Five Newly Designed Modules To Enhance Accuracy, Interpretability, And Operational Scalability. First, The Pathological-Temporal Decomposition Model (PTDM) Leverages Discrete Wavelet Transform To Extract Latent Progression Patterns From Time-series Blood Parameters, Improving Early-stage Detection. Second, The Clinical-Spectrum Weighted SVM Ensemble (CSW-SVM) Introduces Severity-informed Kernel Weighting For Improved Stratification Across Mild To Severe Cases. Third, The Causal-Attention Based Explainable Layer (CAX-EL) Fuses Causal Inference With Attention Networks, Delivering Transparent And Patientspecific Feature Importance Rankings. Fourth, The Patho-Geo-Spatial Outbreak Mapping Model (PGOMM) Integrates Pathological Signals With Geolocation Data Using Graph Anomaly Detection To Forecast Outbreak Clusters. Finally, The Multi-Objective Dengue Outcome Predictor Via HyperFeature Fusion (MOD-HFF) Employs Multi-task Neural Learning To Simultaneously Predict Severity, Hospitalization Likelihood, And Recovery Duration. Together, These Methods Deliver A High-resolution Diagnostic And Forecasting System, Yielding Detection Accuracy Of 91.2%, Severity-wise F1-score Of 0.89, And Outbreak Hotspot Detection With 93.6% Sensitivity. This Work Advances The Frontier In AI-assisted Outbreak Intelligence By Optimizing Temporal Dynamics, Clinical Relev

Published:

09-6-2025

Issue:

Vol. 25 No. 6 (2025)

Page Nos:

400-407

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite