Enriched nonlinear grey compositional model for analyzing multi-trend mixed data and practical applications.

The compositional data are interrelated, and analyzing the evolution of each component is crucial for understanding population dynamics. However, the complex structure and tedious process of modeling pose challenges to the reasonable construction of grey compositional models for analyzing multi-trend mixed data. To address this, a novel enriched nonlinear grey compositional model with global multi-parameter combinatorial optimization is firstly proposed. Secondly, two types of Monte Carlo simulations are designed to validate the performances, modeling characteristics and noise levels of our model. Finally, using the bioenergy power generation structure of China as a case study, the practicability of our approach is verified. The results demonstrate that our model significantly outperforms traditional mainstream models in multi-trend mixed sequences, and the interrelationships among components are effectively verified. Our model not only enriches the methodological base but also broadens the application scope of grey compositional model. • A novel non-linear dynamic GM-Markov compositional model is constructed. • Our model is capable of achieving global multi-parameter combinatorial optimization. • Our model can accurately fit compositional data exhibiting fluctuations and multi-trend. • Our model might be useful in forecasting bioenergy power generation structures. [ABSTRACT FROM AUTHOR]