A Hype Detector based on LSTM, Stock Prices from Market Sentiment and Technical Indicators

The chaotic nature of financial markets has historically posed challenges to traditional time series models like Long Short-Term Memory (LSTM), Autoregressive Integrated Moving Average (ARIMA), and Gated Recurrent Units (GRU), resulting in limited success in accurate stock price predictions. To address these limitations, this paper presents a neural network that integrates market sentiment with stock prices and technical indicators to enhance prediction accuracy. We use sentiment data from Indian financial news websites and historical stock data from Yahoo Finance. Our model combines LSTM for stock prices, dense layers for technical indicators, and an embedding layer with LSTM for sentiment analysis, providing a comprehensive approach to market forecasting. The model attains a scaled Root Mean Squared Error (RMSE) of 0.1010, which corresponds to an RMSE of 1.36 INR when unscaled, against an average stock price of 243.73 INR. These results demonstrate a significant improvement over baseline models that rely solely on stock price inputs and sentiment analysis.