An escape-aware chaos reflection enhanced coati optimization algorithm for Parkinson's disease classification

Parkinson’s disease is a neurodegenerative disorder that can be automatically diagnosed using handwriting. However, due to limitations in data collection devices and scenarios, the feature sets used for Parkinson’s disease diagnosis often face high dimensionality, noise, and redundant features. Addressing these challenges is crucial for improving the accuracy of Parkinson’s disease diagnosis. Feature selection methods, which reduce feature dimensionality and extract more relevant feature sets, are commonly employed to tackle these issues. However, traditional feature selection methods often show limitations when dealing with high-dimensional features. Thus, this paper proposes an Escape-Aware Chaos Reflection Enhanced Coati Optimization Algorithm (ECRCOA) for feature selection. Specifically, a population initialization strategy based on the Fuch chaotic map and quasi-refection-based learning (FQ) is proposed to ensure better diversity and distribution of the initial population, enhancing the algorithm’s global exploration capability. In addition, a new exploration phase integrating the prey escape mechanism (PE) is designed. By simulating prey escape behaviors, the search direction is dynamically adjusted to effectively balance the exploration and exploitation phases and prevent the algorithm from being trapped in the local optimum. The proposed algorithm’s effectiveness is validated on the PaHaW dataset.