Enhanced Vaccine Recommender System to prevent COVID-19 based on Clustering and Classification

Due to the fact that countries are presently dealing with the third wave of COVID-19 pandemic and in present time, the data of vaccines for preventing COVID-19 has triggered massive information, it is vital to create a system that can assist decision-makers and health care practitioners in combating COVID-19 and to combat the problem of vaccine information overload is to provide patients with personalized vaccine recommendations. Because of the ability of recommender systems (RSs) that use Collaborative Filtering (CF) to interpret decision-maker expectations, methodologies, it widely used and direct them towards linked tools that are acceptable to recommend the suitable vaccine for the persons. In this paper, we adopted an Enhanced Vaccine RSs for preventing COVID-19, which is called EVRSs-19. EVRSs-19 face some problems such as sparsity and diversity of vaccines data. To overcome these problems, we adopted two proposals. First, use clustering of K-Means to cluster the persons in several groups according to vaccine types to cope with sparsity of vaccines data. Second, use the K-Nearest Neighbors classifier-depend model of CF to discover neighbors in each vaccine cluster to increase diversity. Evaluating the EVRSs-19 system implemented on vaccines data in two testing using some metrics and the findings of these metrics after running the clustering and classification show that the system of EVRSs-19 has a perfect structure. Such as recall (0.92), precision (0.89), diversity score (9). As the vaccines recommendation list progressed, NDCG and DCG for persons are decreased.