Neural Network and Sentimental Model for Prediction of Stock Trade Value

Abstract

Forecasting and pattern recognition are increasingly important in unpredictable of the stock market. No system can consistently deliver correct predictions; complex machine learning approaches are required. Many research initiatives from numerous disciplines have been carried out to address the difficulties of stock market forecasting. In order to predict stock values, a significant amount of machine learning research has been conducted. Many machine learning techniques have been applied to this form of forecasting, and the results were satisfactory. In this study, we'll utilize web scraping to get all the actual data from the National Stock Exchange (NSE) and Long Short Term Memory (LSTM) Networks with prior data mining techniques to try and forecast the value of the stock market on a certain day. The results of this study show the potential of LSTM Networks for examining historical stock price data and obtaining useful guidance through trend forecasting with the appropriate economic parameters. To determine if a company's stock price is heading upward or lower, should also gather all the most recent commentary from the pertinent websites and apply noise reduction, a classifier, and an algorithm to analyze the sentiment polarity. Using this method, the proposed system represents the current condition of specific stock information.