Blockchain-Secured Recommender System for Special Need Patients Using Deep Learning.

Area of Science:

• Computer Science, Artificial Intelligence, and Health Informatics.
• Focuses on the intersection of recommender systems, deep learning, and data privacy in healthcare.

Background:

• Recommender systems (RSs) are valuable for hospital data management and patient care, especially for those with special needs.
• Existing diet recommender systems struggle to ensure cryptographic data privacy for sensitive patient information.
• The subtle nature of hospital-patient data necessitates robust privacy-preserving techniques.

Purpose of the Study:

• To develop a secure and privacy-preserving diet recommendation system for patients with special needs.
• To integrate a blockchain privacy system (BPS) with deep learning models to protect patient confidentiality.
• To evaluate the performance of various machine and deep learning algorithms within the proposed secure framework.

Main Methods:

• Incorporation of a blockchain privacy system (BPS) into a deep learning framework for diet recommendations.
• Implementation and evaluation of machine learning (ML) and deep learning (DL) algorithms (RNN, Logistic Regression, MLP, LSTM).
• Utilized an Internet of Medical Things (IoMT) dataset comprising 50 patients' data and 1,000 products, analyzed via BPS and encoded features.

Main Results:

• The Long Short-Term Memory (LSTM) deep learning model achieved superior performance in prediction accuracy, precision, F1-measure, and recall within the secured BPS.
• LSTM model attained 97.74% accuracy, with 98% precision, 99% recall, and 99% F1-measure for the allowed class.
• The system demonstrated effectiveness in securing communication channels and enhancing deep learning for personalized dietary recommendations based on patient features.

Conclusions:

• The proposed Blockchain Privacy System (BPS) integrated with deep learning, particularly LSTM, significantly enhances the security and accuracy of diet recommender systems for patients with special needs.
• The system effectively protects patient confidentiality while providing personalized dietary recommendations.
• This novel approach represents a significant advancement over existing literature in privacy-preserving recommender systems for healthcare.