基于混沌离散粒子群优化的约束性多分类模型

Restrictive Multi-classes Classification Model Based on Chaotic Binary Particle Swarm Optimization

约束性多分类问题是在某些工程和生产领域中存在的一类具有特殊约束条件的多分类模式识别问题。针对传统的有监督分类法无法解决约束性多分类问题,提出一种基于混沌离散粒子群优化的约束性多分类模型(CBPSO-RMCM),并将该模型应用于盾构隧道管片选型预测。仿真实验表明,CBPSO-RMCM模型能有效地实现约束性多分类模式识别,并且分类准确率较高。

The problem of restrictive multi-classes classification is a kind of multi-classes pattern recognition subject which contains specific restrictive condition in some engineering and production area. Aiming at the question that the traditional supervised classification method could not solve the problem of restrictive multi-classes classification, this paper proposes a Restrictive Multi-classes Classification Model based on Chaotic Binary Particle Swarm Optimization（CBPSO-RMCM）, and uses this model to predict the selection of segment type in shield tunneling process. Simulation experiment shows that the CBPSO-RMCM model is capable of solving the problem of restrictive multi-classes classification, and has high classification accuracy.