שמור ב:
| Main Authors: | , , , , |
|---|---|
| פורמט: | Recurso digital |
| שפה: | |
| יצא לאור: |
Zenodo
2026
|
| נושאים: | |
| גישה מקוונת: | https://doi.org/10.5281/zenodo.18712930 |
| תגים: |
הוספת תג
אין תגיות, היה/י הראשונ/ה לתייג את הרשומה!
|
תוכן הענינים:
- This project presents a Smart Community Health Monitoring and Early Warning System (SCHM-EWS) designed to detect, classify and provide early warnings for water-borne disease outbreaks at the community/district level. The system integrates crowdsourced symptom reports (public users), field data and sample uploads (ASHA workers), and environmental/wastewater signals with a modular AI pipeline that performs symptom classification, anomaly detection, and short-term outbreak forecasting. The AI layer combines natural language models (fine-tuned transformer-based models) for free-text symptom classification with gradient-boosted trees and LSTM/temporal models for tabular and time-series forecasting[1]. The platform supports role-based access (public / ASHA / government), privacy-preserving data handling, and an operator dashboard for rapid response. Expected outcomes include improved detection lead time, higher sensitivity for localized outbreaks, and actionable alerts for public health teams to prioritize sample collection and interventions. Practical deployment considerations (data quality, connectivity, and ethical safeguards) and evaluation metrics (precision, recall, F1, lead time improvement, and AUC for forecasts) are discussed.