A Novel Approach for Predict Stock Prices in Fusion of Investor's Sentiment and Stock Data

This study employs forecast stock market trends by integrating benchmark sentiment analysis and deep learning techniques. In periods of consistent growth within the stock market, technical analysis is the predominant method for making predictions. Technological prediction involves utilizing stock data obtained through web crawling from” EconomicTimes.com” This data is then used to predict a model, compared to the predictions generated by multiple other models. However, technological prophecy fails to align accurately with stock price trends during significant market volatility. Consequently, benchmark prediction emerges as the most reliable approach for forecasting. The utilization of benchmark prediction techniques enables the anticipation of stock market trends based on the dissemination of information by shareholders through social platforms, particularly in situations of urgency. The proposed approach for benchmark prediction involves extracting stock comments from a specified source and subsequently scrutinizing these comments to identify and filter out any instances of false news. Once the verification process has been conducted on the misinformation, the words that have been deemed to be accurate news undergo preprocessing. Once the authentic news comments have been transformed into textual segments, sentiment analysis is applied. The outcomes of the sentiment analysis are then utilized to generate indicators that reflect fluctuations in sentiment. Ultimately, the ascent and descent indicators are integrated with stock price data to construct a graphical representation, from which the forthcoming trajectory of stock prices is projected based on the curve's pattern. This paper addresses the limitation of the technical aspect in accurately predicting outcomes under exceptional circumstances and significantly enhances the accuracy of stock prediction in real-world scenarios.