HOA Algorithm Optimized L1-Regularized Conformable Fractional Grey System Model

This paper introduces an L1-regularized conformable fractional grey model (CFGM-L1) optimized using the hiking optimization algorithm (HOA). The CFGM-L1 model aims to enhance predictive accuracy by addressing the limitations of traditional grey models in handling high-dimensional and complex data. By incorporating L1 regularization, the model effectively reduces the number of parameters, automatically selects significant features, and prevents overfitting, thereby improving generalization. The HOA is employed to optimize the parameters of the CFGM-L1 model, leveraging a novel metaheuristic approach inspired by hikers' movement patterns to achieve better performance in complex data scenarios. Numerical examples demonstrate the superiority of the CFGM-L1 model over traditional grey models, showcasing significant improvements in prediction accuracy with reduced mean absolute percentage error (MAPE). The results indicate that the proposed approach is a robust and effective method for complex data modeling, with potential applications across various fields.