Dynamic Prediction and Volatility Analysis of the Hang Seng Index based on the VAR-LSTM Model

The Hang Seng Index, as a key index in the Hong Kong financial market, its accurate prediction is of great significance to investors, financial institutions, and policymakers. Given that traditional prediction methods struggle to capture the non-linear characteristics and complex patterns of financial data, this study constructs a VAR-LSTM combined model for predicting the Hang Seng Index. The research selects Hang Seng Index data and multiple macroeconomic indicators over a specific period, screens out key indicators through correlation coefficient analysis, and compares the prediction performances of multiple different LSTM models and the VAR-LSTM model. The results show that the VAR-LSTM model performs remarkably well in fitting data and predicting unknown data, outperforming other models in various error indicators of the training set and the test set. This indicates that the VAR-LSTM combined model has promising prospects in predicting the Hang Seng Index and can provide more accurate prediction information for market participants. However, the financial market is complex and volatile. In the future, it is still necessary to explore methods to optimize the model to further improve the accuracy and reliability of predictions.