A new hybrid method with data-characteristic-driven analysis for artificial intelligence and robotics index return...

Area of Science:

• Financial econometrics
• Artificial intelligence applications
• Robotics market analysis

Background:

• Forecasting returns for the Artificial Intelligence and Robotics (AIRO) Index is vital for financial market stability and AI industry growth.
• Existing methods may not fully capture the complex, time-varying nature of AIRO index returns.
• Investors require reliable references for AI index investments.

Purpose of the Study:

• To propose innovative hybrid methods for forecasting AIRO index returns.
• To account for the multiple structural, time-varying, and nonlinear characteristics of the index.
• To enhance the accuracy and reliability of AI index return predictions.

Main Methods:

• Ensemble Empirical Mode Decomposition (EEMD) for index return decomposition.
• Modified Iterative Cumulative Sum of Squares (ICSS) algorithm for structural breakpoint identification.
• Hybrid forecasting models combining Particle Swarm Optimization-Least Square Support Vector Machine (PSO-LSSVM) and Generalized Autoregressive Conditional Heteroskedasticity (GARCH) type models.

Main Results:

• AIRO index returns exhibit complex structural, time-varying, and nonlinear characteristics.
• The proposed hybrid forecasting method (EEMD-PSO-LSSVM-ICSS-GARCH) demonstrates optimal forecasting performance.
• The model effectively captures the high complexity and mutability of the AIRO index returns.

Conclusions:

• The AIRO index presents significant complexity, necessitating advanced forecasting approaches.
• The novel EEMD-PSO-LSSVM-ICSS-GARCH model offers superior predictive accuracy for AIRO index returns.
• This research provides a more reliable tool for investors in the artificial intelligence and robotics sectors.