Healthcare Recommender Systems Enhanced by Learnable Discrete Wavelet (LDW) Pooling in Deep Learning

Recommender systems have become increasingly important in the healthcare sector due to their vital role in predicting various health-related information for both patients and doctors. The primary goal of this system is to guarantee the timely availability of essential information while upholding its quality, reliability, and authentication. The primary emphasis in conventional CNN design lies in the convolutional layers and the activation functions they employ. The Learning Discrete Wavelet Pooling (LDW-Pooling) technique offers a universal alternative to standard pooling operations, resulting in enhanced accuracy and efficiency in feature extraction. This research paper introduces an Intelligent Health Recommender System (HRS) that utilizes the Discrete Wavelet Pooling (LDW-Pooling)-Convolutional Neural Network (CNN) technique. The performance evaluation of this intelligent recommendation system is conducted using a well-known Multi disease dataset (Heart, Liver, and Kidney). Evaluation based on different parameter values demonstrates that the proposed deep learning method (LDW-Pooling) outperforms other approaches by producing fewer errors. The proposed system achieves an impressive accuracy rate of 98.1%, surpassing the performance of existing deep machine-learning methods. This indicates that our system is highly suitable for multidisciplinary disease prediction and recommendation.