基于差异协作策略和稳定生长策略的青蒿素优化算法及其在高维医学特征选择问题的应用

Artemisinin Optimization Based on Differential Collaboration and Stable Growth Strategies and Its Application in High-Dimensional Medical Feature Selection Problems

摘要: 高维医学数据普遍存在特征维度高、样本量小、冗余与噪声干扰等特征，易造成计算负担过重、模型过拟合及诊断精度下降等问题，特征选择是解决该问题的关键技术。元启发式算法在特征选择领域应用广泛，但原始青蒿素优化算法在处理高维医学特征选择任务时，易出现收敛停滞现象，导致特征筛选效率偏低。为此，文章提出一种融合差异协作策略与稳定生长策略的改进青蒿素优化算法，通过个体间差异协作实现有效信息交互与协同寻优，利用稳定生长策略维持种群多样性，协同提升算法的全局搜索与局部优化能力，进而增强其在高维医学数据中的特征选择效率与筛选效果，为医学数据挖掘与智能诊疗提供更高效的特征优化方法。

Abstract: High-dimensional medical data generally exhibits characteristics such as high feature dimensionality, small sample size, and interference from redundancy and noise, which easily lead to problems including excessive computational burden, model overfitting, and reduced diagnostic accuracy. Feature selection serves as a key technique to address these issues. Although metaheuristic algorithms have been widely applied in the field of feature selection, the original Artemisinin Optimization Algorithm is prone to convergence stagnation when handling high-dimensional medical feature selection tasks, resulting in low efficiency of feature screening. To this end, this paper proposes an improved Artemisinin Optimization Algorithm integrating a differential collaboration strategy and a stable growth strategy. It realizes effective information interaction and collaborative optimization through differential collaboration among individuals, maintains population diversity via the stable growth strategy, and synergistically enhances the global search and local optimization capabilities of the algorithm. Consequently, it improves the efficiency and performance of feature selection on high-dimensional medical data, providing a more efficient feature optimization method for medical data mining and intelligent diagnosis and treatment.