Design and Development of Artificial Intelligence Framework to Forecast the Security Index Direction and Value in Fusion with Sentiment Analysis of Financial News

The domain of stock price prediction is extensively researched owing to its complex data structure and numerous influential factors. In the current epoch, many modern financial applications demonstrate non-linear and uncertain characteristics that exhibit temporal variability. As a result, there has been a notable rise in the need for resolutions to highly non-linear and time-variant issues. External factors, including public sentiment and political events, can influence stock market trends. The primary aim of this research is to propose a novel framework for predicting the blue-chip stock price in the future by combining historical stock data corresponding to the financial news published regarding stock in the newspaper. The initial step involves integrating sentiment and situational features into a machine-learning model. This integration aims to evaluate public sentiments’ influence on algorithms’ predictive precision thirty days ahead. Furthermore, regression models are utilized to analyze the inter-dependencies that exist among companies. To facilitate an experimental analysis, the researchers acquired historical data from the stock market through reputable sources such as the National Stock Exchange-India and collected news data about stock from the financial newspaper “EconomicTimes”. A series of machine learning models were tested to facilitate the classification of news sentiments of the stocks, followed by the fusion of resultant polarities with the stock’s historical data on the standard column “DATE” and then tested series of deep learning models to observe the impact of news on stock price and by taking reference of this, the authors successfully predicted the stock price on a random day. The findings unequivocally indicate that the proposed research design produces better results than other state-of-the-art models.