A Novel Feature Selection Method for High-Dimensional Biomedical Data Based on an Improved Binary Clonal Flower Pollination Algorithm
A Novel Feature Selection Method for High-Dimensional Biomedical Data Based on an Improved Binary Clonal Flower Pollination Algorithm
复制标题
基于改进二元克隆花授粉算法的高维生物医学数据特征选择新方法
DOI：
10.1159/000501652
复制
发表时间：
2019-09-01
期刊：
影响因子：
1.8
通讯作者：
Luo, Junwei
中科院分区：
文献类型：
--
作者：
Yan, Chaokun;Ma, Jingjing;Luo, Junwei
在生物医学领域，大量的生物和临床数据迅速积累，这些数据可被分析以强调对高危患者的评估并改进诊断。然而，生物医学数据分析面临的一个主要挑战是所谓的“维度灾难”。针对这一问题，提出了一种基于改进的二进制克隆花朵授粉算法的新型特征选择方法，以消除不必要的特征并确保对疾病进行高精度分类。采用了绝对平衡组策略和自适应高斯变异，这可以增加种群的多样性并提高搜索性能。使用KNN分类器来评估分类准确性。在六个公开可用的高维生物医学数据集上的大量实验结果表明，所提出的方法能够获得较高的分类准确性，并且优于其他最先进的方法。
In the biomedical field, large amounts of biological and clinical data have been accumulated rapidly, which can be analyzed to emphasize the assessment of at-risk patients and improve diagnosis. However, a major challenge encountered associated with biomedical data analysis is the so-called “curse of dimensionality.” For this issue, a novel feature selection method based on an improved binary clonal flower pollination algorithm is proposed to eliminate unnecessary features and ensure a highly accurate classification of disease. The absolute balance group strategy and adaptive Gaussian mutation are adopted, which can increase the diversity of the population and improve the search performance. The KNN classifier is used to evaluate the classification accuracy. Extensive experimental results in six, publicly available, high-dimensional, biomedical datasets show that the proposed method can obtain high classification accuracy and outperforms other state-of-the-art methods.