Developing Hybrid Deep Learning Models for Stock Price Prediction Using Enhanced Twitter Sentiment Score and Technical Indicators

In recent years, there has been growing interest in using deep learning methods to improve the accuracy of stock price prediction, which has always been challenging due to the unpredictable nature of the market. This paper introduces two new hybrid deep learning-based models, named “En-Tweet-Deep-SMF” and “En-Tweet-Hib-SMF,” that combine effective strategies to enhance stock price prediction accuracy. These strategies involve enhancing Twitter sentiment scores using an enhanced model and utilizing potent technical indicators. The “En-Tweet-Deep-SMF” model employs a gated recurrent unit, while the “En-Tweet-Hib-SMF” model uses the convolutional neural network-bidirectional long-short term memory hybrid deep learning-based model. Additionally, kernel principal component analysis is utilized to reduce the dataset dimensionality. These models can capture both quantitative and qualitative factors that can influence stock prices, making them more accurate and robust than traditional methods. The proposed models have the potential to adapt and learn from new data and trends, providing traders, investors, and financial analysts with a valuable tool to make informed decisions and mitigate risks in the stock market. Experimental results indicate that these models outperform several state-of-the-art models, demonstrating their effectiveness and potential practical applications in the financial industry.