Financial Decision-Making AI-Framework to Predict Stock Price Using LSTM Algorithm and NLP-Driven Sentiment Analysis Model

Predicting stock market fluctuation is a crucial field in artificial intelligence – AI and financial technology – FinTech research due to its significance and implications for investors and their investment strategies. It is evident from recent research studies that sentiment analysis has a significant impact on the stock market. This paper presents a stock price prediction model to conduct experiments using financial time series and financial news datasets. This study has retrieved sentiment data from News API and processed it to quantify market sentiment. Using natural language processing (NLP) techniques, sentiment scores have been evaluated as positive or negative news and analysed sentiments have been incorporated with historical stock price data retrieved from NASDAQ. Data has been pre-processed to normalize the datetime index and merge the closing price with sentiments to ensure the consistency and suitability of the data for training the prediction model. A multi-layer LSTM neural network model has been identified as a suitable prediction model employed on stock prices and sentiment dynamics of the financial market for highly accurate stock price prediction. The multi-layer LSTM model has been fine-tuned using different parameters such as different neuron layers, epochs and batch size. Prediction results and model accuracy have been evaluated using the following metrics: root mean square error – RMSE, mean absolute percentage error – MAPE, and R-squared. The proposed model enhances accuracy in predicting short-term stock price trends for millennials as they are mostly aggressive investors who would like to make a profit in a shorter phase. Integrating sentiment data improved the performance of prediction models, highlighting the critical role of stakeholders' sentiment in stock market performance. The results of this study are a valuable contribution to the growing field of the AI-driven financial sector, demonstrating the viability of integrating NLP-driven sentiment analysis with deep learning to make more informed investment decisions.