Stock Price Prediction Using Optimal Network Based Twitter Sentiment Analysis

Abstract

In recent times, stock price prediction helps to determine the future stock prices of any financial exchange. Accurate forecasting of stock prices can result in huge profits to the investors. The prediction of stock market is a tedious process which involves different factors such as politics, economic growth, interest rate, etc. The recent development of social networking sites enables the investors to discuss the stock market details such as profit, future stock prices, etc. The proper identification of sentiments posted by the investors in social media can be utilized for predicting the upcoming stock prices. With this motivation, this paper focuses on the design of effective stock price prediction using dragonfly algorithm (DFA) based deep belief network (DBN) model. The DFA-DBN technique aims to properly determine the sentiments of the investors from Twitter data and forecast future stock prices. From Twitter data, the DFA-DBN technique attempts to accurately determine the sentiments of investors, as well as predict future stock prices. For accurate stock price prediction, the proposed DFA-DBN model includes the development of a DBN model. The proposed DFA-DBN model involves the design of DBN model for accurate prediction of stock prices. Besides, the hyperparameter tuning of the DBN technique is performed by utilize of DFA and thereby boosts the overall prediction performance. For validating the supremacy of the DFA-DBN model, a comprehensive experimental analysis takes place and the results demonstrate the accurate prediction of stock prices. A predicted DFA-DBN algorithm with a higher accuracy of 94.97 percent is available. On the basis of the data in the tables and figures above, the DFA-DBN approach has been demonstrated to be an effective instrument for anticipating stock price fluctuations.