Recurrent Neural Networks for Feature Extraction from Dengue Fever

Area of Science:

• Public Health
• Computational Biology
• Data Science

Background:

• Dengue fever outbreaks pose a significant global health challenge, necessitating effective control strategies.
• Accurate early dengue fever projections are vital due to the lack of specific treatments and universal vaccines.
• Neural networks offer promising advancements in public health applications, including disease surveillance.

Purpose of the Study:

• To develop and evaluate a deep learning model for enhanced feature extraction from dengue fever text datasets.
• To assess the efficacy of the Random Forest (RF) algorithm in identifying key dengue fever characteristics.
• To improve the accuracy and reduce errors in dengue fever data analysis for better outbreak prediction.

Main Methods:

• Data collection and preprocessing of dengue fever-related text datasets.
• Implementation of a deep learning model incorporating the Random Forest (RF) algorithm for feature extraction.
• Comparative analysis of the proposed RF-based feature extraction method against existing techniques.

Main Results:

• The proposed RF-based deep learning model demonstrated a significant improvement in feature extraction accuracy, exceeding existing methods by over 12%.
• The study achieved a 10% reduction in feature extraction errors compared to other conventional methods.
• Simulation results validated the model's effectiveness in extracting relevant dengue fever features from text data.

Conclusions:

• The developed deep learning model utilizing Random Forest (RF) offers a superior approach for dengue fever feature extraction from text.
• This enhanced feature extraction capability is crucial for improving early dengue fever outbreak prediction and control strategies.
• The model's high accuracy and error reduction highlight its potential for advancing vector-borne disease management.