LC-CA-FF-Mamba in the Field of Securities Market Prediction

A Mamba-based time series forecasting algorithm is improved to address the lack of accuracy as well as the inefficiency of traditional forecasting methods for time series in securities market prices. The algorithm uses a novel state space model (SSM) with linear time complexity, which is improved on the channel attention-feature fusion-mamba (CA-FF-Mamba).By adding UnetTSF, a long time series prediction module with linear complexity, a new model of linear complexity -channel attention-feature fusion-mamba (LC-CA-FF-Mamba) is proposed which, by deeply mining the intrinsic patterns of historical transaction data, in Reducing the reliance on manual feature engineering and complex data preprocessing, the model achieves high-precision prediction of future securities market price movements. Multiple controlled experiments show that the LC-CA-FF-Mamba model significantly outperforms existing benchmark methods in key metrics in the prediction task for multiple securities market species. This improved forecasting capability can provide reliable technical support for institutional investors to construct quantitative strategies and optimize investment returns with controlled risks. The results of this research not only validate the effectiveness of Mamba architecture in financial time series analysis but also provide a new technical path for the development of intelligent investment systems.