论文部分内容阅读
针对模糊C-均值聚类算法存在对初始聚类中心敏感和聚类目标函数容易陷入局部最优的问题,提出了1种基于混沌差分进化模糊C-均值聚类的多模型建模方法。该方法采用混沌差分进化算法对模糊C-均值聚类的目标函数进行全局寻优,能有效的解决上述问题。将该方法应用于双酚A生产过程的质量指标软测量建模,仿真结果表明了该算法的有效性。
Aiming at the problem that the fuzzy C-means clustering algorithm is sensitive to the initial clustering center and the clustering objective function tends to fall into the local optimum, a multi-model modeling method based on the chaos differential evolution fuzzy C-means clustering is proposed. This method uses the chaos differential evolution algorithm to globally search the objective function of fuzzy C-means clustering, which can effectively solve the above problems. The method was applied to the soft index modeling of the quality index of BPA production process. The simulation results show the effectiveness of the algorithm.