<p>Creating a stock market forecasting model is challenging for any emerging market such as the Indian stock market. This is mainly due to lesser research work and most of these research works that involved the past prices to predict future prices and also the sentiment of investors from the economic factors such as major Western stock market movements, fluctuation in commodity prices in the global market, war escalations such as the Middle East crisis and Ukraine war. Central bank actions are often influencing the stock price movement in emerging stock markets. The existing stock price forecasting prediction models lack the consideration of investor’s sentiment from various economic factors that influence price movement. To overcome this challenge, we have proposed a hybrid forecasting model to combine the sentiment analysis data of economic information using natural language understanding application with the past prices of stock or index in a deep learning multivariate long short-term memory (LSTM) model to predict the future open and close prices. The result suggests that the sentiment analysis data on economic information have a strong influence on the stock price movement and the multivariate LSTM deep learning model achieved the mean absolute percentage error (MAPE) of 0.5676%, and the coefficient for the determinant (R2) of 0.9743 for the predicted close price. The accuracy of prediction of the stock or index prices through the LSTM deep learning models improves using the natural language understanding (NLU)-based sentiment analysis score of economic data. Based on our findings, future research can be performed on other market in which sentiment of investor or customers on various events can influence the price movements. Apart from the financial instruments, this approach can be used in consumer-driven product forecasting models.</p>