A Comprehensive Analysis of Social Media Data-Based Election Prediction: Research Obstacles and Future Prospects

The introduction and recognition of modern-day social media (SM) which include Facebook, Twitter, and Instagram have converted the way politicians talk with their citizens and marketing campaign for elections. The inherent talents of SM, together with the potential to manner massive quantities of statistics in actual time, have brought about the emergence of a brand new discipline of research that uses SM information to predict election outcomes. Although many studies were carried out during the last decade, their results are exceptionally arguable and widely debated. In this context, the goal of this article is to explore and summarize how studies on predictive preference based totally on SM facts has developed on the grounds that its inception, thereby reflecting each the state of the artwork, practices, and opportunities for studies in this location. In the context of election research, we have described the primary techniques and features of a hit research, the principle strengths and issues, and as compared our consequences with previous reviews, concerning technique, amount, and satisfactory of guides. We diagnosed and analysed eighty three relevant research, and identified problems in numerous areas, inclusive of process, sampling, layout, choice-making, and medical rigor. Key conclusions encompass the low success rate in the use of the most commonplace methods, Twitter quantity and sentiment analysis, and the better outcomes of more recent strategies, which includes educated regression methods, in traditional studies. Finally, a angle for future research at the coordination of trends in the definition, modelling, and evaluation processes is also mentioned, which, among different things, demonstrates the want to better recognize the software of contemporary system learning methods.