A Hybrid Prediction Model Based on Decomposition-Integration for Foodborne Disease Risks

Area of Science:

• Public Health
• Epidemiology
• Data Science

Background:

• Foodborne diseases (FBDs) pose significant global economic and health challenges.
• Accurate prediction of FBD risk trends is a critical public health objective.
• Existing prediction methods require enhancement for improved accuracy and reliability.

Purpose of the Study:

• To develop and validate a novel risk prediction model for foodborne diseases.
• To improve the accuracy and reliability of FBD risk forecasting.
• To provide data-driven support for food safety management and policy.

Main Methods:

• Utilized a decomposition-integration technique on FBD surveillance data (2019-2023, Wuxi).
• Employed complete ensemble empirical mode decomposition with adaptive noise to decompose FBD risk data into intrinsic mode functions (IMFs).
• Reconstructed IMFs using sample entropy and analyzed time dependence with a temporal convolution network-long short-term memory (TCN-LSTM) model.

Main Results:

• The proposed TCN-LSTM model demonstrated superior prediction accuracy for FBD risks.
• Achieved an average root mean square error of 5.349 and mean absolute error of 3.819.
• Showcased at least a 40% improvement in prediction accuracy compared to standalone LSTM models.

Conclusions:

• The developed decomposition-integration TCN-LSTM model offers a significant advancement in FBD risk prediction.
• The model provides valuable data support for effective food safety management and policy formulation.
• Enhanced FBD risk prediction capabilities enable more accurate and timely public health early warnings.