Artificial Intelligence-Driven Financial Market Trend Forecasting and Quantitative Portfolio Dynamic Adjustment Strategy Design

This paper examines the efficacy of artificial intelligence predictive models to analyse the behaviour of the financial market and to manage the corresponding dynamic allocation of portfolios. Collecting BlackRock’s stock data (2013-2023) and macroeconomic fundamentals of GDP growth and interest rates, the performance of ARIMA and LSTM models is compared. It is also found that LSTM has a better predictive performance compared to ARIMA when compared in terms of RMSE (10.26) and MAE (7.47). Macro quantitative data regression shows that macro fundamentals are not related to near-term portfolio returns, highlighting micro market characteristics. The discussed results explain the benefits of utilising AI-based models for improving predictive results as well as using actual portfolio changes.