A Sentiment Analysis of Political Tweets for the Purpose of Public Opinion

Abstract

For example, forecasting the outcomes of national and municipal elections requires surveying public opinion. This is a typical example of a job in political analysis. But surveys may be costly to do, and if the sample isn't representative of the community, the findings might be skewed. We provide two novel approaches to using the findings of tweet sentiment analysis in survey research. Using only the results of the tweet sentiment analysis, our initial technique outperformed many well-known algorithms. We also present a new hybrid method that estimates election outcomes using public opinion surveys and Twitter. This approach is used to estimate the outcome of the 2023 Greek national election and allows for more accurate, frequent, and cost-effective public opinion estimates. New opportunities for public opinion estimate using social media platforms were opened up by our technique, which showed less variance from the actual election outcomes than traditional public opinion surveys.