An Integrated Framework of Machine Learning and Evolutionary Computation for Prediction of Stock Exchange Based on Covid-19 News

Due to its importance in the individual and national economies, stock market prediction is an important study subject. A stock market’s success is based on the rise or fall of industrial, financial, medical, local, and global stock prices in a certain region and globally. Electronic news and public opinion impact stock markets. Since the Covid-19 epidemic, global stock prices have fluctuated because to economic uncertainty. Social media has been used to spread pandemic news, comments, and forecasts. Such news has affected global stock marketplaces, which are prone to polarity, by affecting investors’ decision-making and changing the perspective of stock-interested people. This paper proposes a methodology to examine the influence of Covid-19 internet news data on stock market performance. News data comes from news.pk and internet news sites, while stock data is from Yahoo Finance. Text characteristics are retrieved from news information using TF-IDF, and stock-related features are produced for stock value prediction. A hybrid approach that combines evolutionary algorithms such as Genetic Algorithm, Harris hawks Optimization, and Particle Swarm Optimization with various Machine Learning and Deep Learning-based models. The empirical analysis-based results reveal the proposed hybrid model outperforms traditional ML models using standard performance evaluation measures.