Introducing a Novel Method to Identify the Future Trend of Nikkei 225 Stock Price in Order to Reduce Investment Risk

Stock market forecasting is a challenging task, as many factors influence the course of the stock market, from economic and political to social events. The challenge, however, underlines the immense importance of basically accurate models for guiding investment decisions. Most of the traditional methods applied in stock market forecasting show inefficiency in capturing the dynamic and nonlinear nature of the market as major setbacks. This study proposes a new incorporation of hyperparameter optimization algorithms into machine learning techniques, such as Genetic Algorithms, Battle Royale Optimization, and Grey Wolf Optimization, for stock price prediction. In an effort to raise the prediction performance of a Multilayer perceptron model by optimizing its hyperparameters, this approach was designed to overcome the difficulties in stock market forecasting. The empirical results showed clearly that the optimized GWO-MLP model outperformed other models with a Coefficient of determination of 0.988, which corresponds to a Root Mean Squared Error of 126.14, Mean Absolute Percentage Error of 0.35%, and Mean Absolute Error of 97.73. These findings point to the efficiency and reliability of the GWO-MLP model in predicting future stock prices as a useful tool for investors. The biggest datasets can be treated in real time by the approach, empowered by recursive neural networks and Grey Wolf Optimization, providing insight with deep market predictions and helping to identify lucrative investment opportunities.