摘要
在乳腺癌识别方法中,支持向量机(SVM)凭借良好的精度和鲁棒性已经取得了不错的预测结果,但是SVM中惩罚系数c和核函数参数g选取的不同也会在一定程度上影响着算法的泛化精度。为了提高SVM的识别性能,提出了一种将改进鲸鱼优化算法(MWOA)和SVM结合的模型。利用MWOA迭代寻优能力对SVM参数进行调整,并以最优化的参数组合对样本数据进行分类识别。为了证明该方法的有效性,应用威斯康辛乳腺癌数据集进行实验并与现有方法进行对比。仿真结果表明,MWOA-SVM与BP神经网络、传统SVM、PSO-SVM、及WOA-SVM 4种方法相比,具备更好的识别性能。
In breast cancer identification methods,support vector machine(SVM)has achieved good prediction results with good accuracy and robustness,However,the difference of penalty coefficient and kernel parameter selection in SVM will also affect the generalization accuracy of the algorithm to some extent.In order to improve the recognition performance of SVM,a model combining modified whale optimization algorithm(MWOA)and SVM was proposed.The parameters of SVM are adjusted by using MWOA iterative optimization capability.Finally,the optimal combination of parameters is used to classify and recognize the sample data.In order to prove the validity of this method,The Wisconsin breast cancer data set was used for experiment and compared with the existing methods.The simulation results show that,compared with BP neural network,traditional SVM,PSO-SVM and WOA-SVM,MWOA-SVM has better recognition performance.
作者
冯云霞
范琳琳
Feng Yunxia;Fan Linlin(School of Information Science and Technology,Qingdao University of Science and Technology,Qingdao 266100,China)
出处
《电子测量技术》
2020年第7期88-92,共5页
Electronic Measurement Technology
基金
国家自然科学基金(61572268、61303193)
山东省重点研发计划项目(2017GSF18110、2018GGX101029)资助
关键词
辅助诊断
改进鲸鱼优化算法
支持向量机
参数优化
分类识别
auxiliary diagnosis
modified whale optimization algorithm
SVM
parameter optimization
cassification identify
