Improving M-Health applications recommendation using fine-tune graph neural networks

The mobile industry is one of the skyrocketing industries of the modern age. Many people rely on using mobile applications in their daily lives. Healthcare facilities are not uniformly available everywhere. To get quick help in some situations, people consult Mobile Health (M-Health) applications. On different mobile app stores, M-Health applications are growing rapidly. Users often face problems in finding the most appropriate application. App descriptions, ratings, and reviews are the only sources to find users’ reviews about certain applications. Users read the text to decide whether to use or skip the application. A sentiment analysis can be used to give the users a fast assessment. Clients’ experiences may also be made better through recommender systems, which can make it easier to find applications that are relevant to them. We propose a novel framework that employs a Graph Neural Network (GNNs) algorithm to develop a recommender system for M-Health applications. The proposed system will help the users to choose the most appropriate application from the entire list. This will also help the software communities to make innovations in the current application by providing users feedback. In order to validate the model performance, the proposed work is tested on custom-built and benchmark datasets, which achieved 97.5% and 98.3% respectively. The proposed work outperforms closely related state-of-the-art works.