LSTM-based Financial Trend Prediction: Exploring the Use of Trend and Momentum Indicators

The ability to forecast financial time series is critical for informed investment strategies and proactive risk control. This paper offers a proposed LSTM-based model for the Classification of the Stock Market as Bullish, Neutral, or Bearish with some Technical Indicators such as Exponential Moving Averages (EMAs), Moving Average Convergence Divergence (MACD), and Relative Strength Index (RSI). The incorporation of several features in feature engineering assists in recognizing significant patterns, and LSTM layers capture The interconnected patterns in stock price fluctuations. The Training and Testing of the model employed a five-year dataset of historical stock prices. The experimental results proved that the model attained an impressive 98% accuracy in predicting market trends reliably. Confusion Matrices and Classification Reports validated its performance over various market conditions. This work highlighted the importance of technical indicators in the prediction process and provided a robust framework for financial time series analysis. The results emphasize the promise of deep learning methods, particularly LSTM networks, to improve forecasting reliability in financial markets.