Evaluation of people’s sentiments towards Covid-۱۹ vaccine side effects based on artificial intelligence methods

Background and aims: Acquiring information about people’s attitudes and sentiments towardsvaccination is an important issue to monitor the acceptance of vaccines among people, whichcauses a more accurate assessment of the community’s needs, which may lead to faster modificationof vaccine properties or vaccination approaches. An important example is the assessmentof sentiments towards covid-۱۹ vaccines that led to better decision-making for governments.Now social media has become a global platform for sharing opinions and talking about varioustopics, making it get used as a data source for extracting people’s attitudes. Studying and dividingattitudes in such a way that has the least theoretical mistake is done by artificial intelligence (AI).In related studies, the positive and negative sentiments towards the vaccination of Covid-۱۹ havebeen discussed, but in this study, for the first time, we systematically review to what extent sideeffects can affect the desire and attitude of people towards the vaccines.Method: We searched the related articles in three databases: PubMed, Web of Science, and Scopus.The keywords for the search were related to artificial intelligence, side effects, and covid-۱۹vaccines.۲۰۳ articles were extracted from databases, among them ۱۳ articles were related to ourwork. The exclusion criteria for article screening were Congress, non-English articles, and articlesincluding terms related to vaccine discovery, etc.Results: According to our investigation, both in the global evaluations and in the evaluation ofa country, the positive attitude of the people towards the vaccine was higher than the negativeattitude or was almost equal.In all evaluations, concerns and negative sentiments about the side effects of the vaccine made upa high percentage of people’s negative attitudes toward the vaccine. The results of ۱۹۲ countriesare in very close agreement with the results at the global level. Attitudes toward some vaccines,particularly the AstraZeneca COVID-۱۹ vaccine, have changed drastically over time. Latent DirichletAllocation (LDA) is the most applied method based on machine learning to form the topicsof sentiment analysis of text mining results, after that, we can refer to Valence Aware Dictionaryfor Sentiment Reasoning (VADER) which is frequently used for scoring the words in the samecluster. The results were variable based on the number of data, and with the increase in the numberof data, the results of different studies were closer.Conclusion: We realized that the side effects of the vaccine were their main obstacle to receivingthe Covid-۱۹ vaccines. So, it appears that the best approach to increase people’s participation inreceiving the vaccine is to increase information about the side effects of the vaccine. It is alsorecommended that future studies address the role of social media in inducing negative opinionsbecause it is possible that some side effects reported by people on social media are only influencedby the declaration and sentiments of others.

کلیدواژه ها:

artificial intelligence ، machine learning ، Covid-۱۹ vaccine ، sentiment ، side effect ، attitude