Hybrid Intelligence for Infectious Disease Forecasting: A Big Data Approach to Predictive Modelling and Risk Pattern Discovery
Year: 2026
Type: article
Abstract: The fast extent of infectious diseases like COVID-19, Influenza, Dengue, and Tuberculosis pretends major challenges to public health systems, exclusively when analysing huge patient datasets. This paper explains a novel hybrid machine learning framework designed for precise infectious disease prediction and critical symptom analysis in a big data environment. The proposed system incorporates Random Forest (RF) for feature selection, K-Means clustering for patient segmentation, and Gradient Boosting (GB) for final classification. The model is trained and validated on a large-scale electronic health record (EHR) dataset covering over 10000 patient records with distinct symptoms, demographics, and clinical attributes. An important contribution of this work is the integration of age-dependent symptom severity analysis utilizing correlation metrics and decision tree-based rule mining. The hybrid model is attained a superior accuracy, outperforming traditional classifiers such as SVM, Decision Tree, and Logistic Regression across metrics like precision, recall, F1 score, and AUC-ROC. Statistical significance testing is further endorsed the enhanced performance. Clustering analysis is revealed distinct patient risk profiles, and feature importance ranking exposed critical symptoms such as breathlessness, chest pain, and comorbidities in elderly populations. This work illustrates the ability of combining ensemble and unsupervised learning in big data setup, delivering an interpretable and scalable solution for infectious disease prediction and clinical decision support.
Authors: Kolhe Shilpa Dattatraya, Ankur Khare
Institution: AISECT University
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W7161602016