一种优化的SVM竹类属种识别方法

An Optimized AS-PO-SVM Classification Model for the Identification of Bamboo Species

提出了一种优化的AS-PO-SVM分类模型,用于解决竹种属的分类问题。AS-PO-SVM是一种基于属性选择(AS)和参数优化(PO)的支持向量机(SVM)分类模型。先用UCI公开数据集验证了AS-PO-SVM模型的分类性能,再将模型应用于由簕竹属、牡竹属、刚竹属和玉山竹属共46个竹种样本构建的Bamboo数据集上。实验结果显示AS-PO-SVM模型在Bamboo数据集上分类准确率达到95.65%,是一种有效的竹种分类模型。

An optimized AS-PO-SVM classification model was proposed and applied in the classification of bamboo plants at the level of genus.AS-PO-SVM is a Support Vector Machines（SVM） classification model based on the attribute selection（AS） and parameter optimization（PO）.The classification ability of AS-PO-SVM model was firstly verified by UCI open data set,and then the model was used in the classification of 46 bamboo species from Bambusa,Dendrocalamus,Phyllostachys and Yushania genus from a bamboo data set.The results showed that classification accuracy of bamboo data set by AS-PO-SVM could attain 95.65%,which suggested that the model is an effective tool for the classification of bamboo plants.