基于双重扰动与改进教与学优化算法的DNA微阵列集成分类

An Ensemble Method Based on Double Disturbances and Improved TLBO for Classifying DNA Microarray

针对DNA微阵列的高维、小样本及高冗余等特点,提出了一种新的集成分类方法.基于bootstrap技术的样本扰动和kruskalwallis与邻域互信息的特征扰动训练多个具有较大差异性和较高准确性的基分类器;针对教与学优化算法易陷入局部最优、优化精度不高和收敛速度较慢等不足,从"教"与"自学"过程入手,设计了一种改进的教与学优化算法实现基分类器的选择性集成,并用于DNA微阵列分类.仿真实验结果表明:该方法在分类精度、集成规模、稳定性等方面具有较强的优势.

In view of the characteristics of high dimensional,small sample and high redundancy,an ensemble method for classifying microarray data is proposed.The sample disturbance based on bootstrap and feature disturbance based on kruskalwallis and neighborhood mutual information are used to train multiple base classifiers in order to improve the diversity among the base classifiers and individual precision.An improved TLBO is designed to realize the selective ensemble from two aspects of the"teaching"and"self-learning"process.Simulation results show that the proposed method has strong advantages in terms of classification accuracy,ensemble size and stability.