Abstract

Model specification is a crucial aspect of structural equation modeling (SEM), since a misspecified model may lead to biased parameter estimation and result in inaccurate conclusions. We propose the Hybrid Ant Colony Optimization Algorithm (hACO), an improved metaheuristic algorithm to conduct model specification searches in SEM. This data mining algorithm combines aspects of the Ant Colony Optimization algorithm with the Tabu search algorithm to increase both accuracy and efficiency. A Monte Carlo simulation study showed that the hACO algorithm provided accurate and efficient SEM specification searches across all designed simulation conditions. The hACO algorithm can help applied researchers conduct specification searches while avoiding potential model misspecifications.

摘要

模型规范是结构方程建模 (SEM) 的一个重要方面，因为错误指定的模型可能会导致参数估计有偏差并导致结论不准确。我们提出了混合蚁群优化算法 (hACO)，这是一种改进的元启发式算法，用于在 SEM 中进行模型规范搜索。该数据挖掘算法将蚁群优化算法的各个方面与禁忌搜索算法相结合，以提高准确性和效率。Monte Carlo 模拟研究表明，hACO 算法在所有设计的模拟条件下提供了准确和高效的 SEM 规范搜索。hACO 算法可以帮助应用研究人员进行规格搜索，同时避免潜在的模型错误规格。