[1]张 颖,窦一峰.基于糖尿病性视网膜病变数据集的支持向量机优化算法比较[J].医学信息,2021,34(02):23-27.[doi:10.3969/j.issn.1006-1959.2021.02.007]
　ZHANG Ying,DOU Yi-feng.Comparison of Support Vector Machine Optimization Algorithms Based on Diabetic Retinopathy Data Set[J].Medical Information,2021,34(02):23-27.[doi:10.3969/j.issn.1006-1959.2021.02.007]
点击复制

基于糖尿病性视网膜病变数据集的支持向量机优化算法比较()

医学信息[ISSN:1006-1959/CN:61-1278/R]

卷:
34卷
期数:
2021年02期
页码:
23-27
栏目:
出版日期:
2021-01-15

文章信息/Info

Title:
Comparison of Support Vector Machine Optimization Algorithms Based on Diabetic Retinopathy Data Set
文章编号:
1006-1959（2021）02-0023-05
作者:
张 颖; 窦一峰
（天津市宝坻区人民医院泌尿外科1，网络信息中心2，天津 301800）
Author(s):
ZHANG Ying; DOU Yi-feng
(Urology Surgery1,Network Information Cente2,Tianjin Baodi Hospital,Tianjin 301800,China)
关键词:
支持向量机; 分类; 参数优化; 预测; 糖尿病性视网膜病变数据
Keywords:
Support vector machine; Classification; Parameter optimization; Prediction; Diabetic retinopathy data
分类号:
TP3
DOI:
10.3969/j.issn.1006-1959.2021.02.007
文献标志码:
B
摘要:
目的 探索支持向量机及其改进算法在糖尿病性视网膜病变数据集上的分类效果，以期更好的探究机器学习方法在辅助医学诊断上的应用价值。方法 以糖尿病性视网膜病变数据集为例，建立支持向量机模型，利用优化思想调整算法参数，运用多指标为所建立的模型对糖尿病性视网膜病变的预测效果进行评价分析。结果 在训练数据占比为50%和60%上，粒子群改进的支持向量机模型在精确率、总精度、马修斯相关系数、F1得分、分类正确率和AUC值等6项指标中均取得了最优结果，在70%的指标上取得了5项指标最优，基于遗传算法改进的模型在80%的训练数据占比上有3项指标最优，基于网格搜索优化的算法在90%的训练数据占比上取得了6项最优结果。结论 改进的支持向量机模型在糖尿病性视网膜病变预测分析中具有较好的应用价值，能够为疾病诊断提供参考依据。
Abstract:
Objective To explore the classification effect of support vector machine and its improved algorithm on medical data set in order to better explore the application value of machine learning method in auxiliary medical diagnosis.Methods The data set of diabetic retinopathy was taken as an example to establish a support vector machine model. The parameters of the algorithm were adjusted with the optimization idea, and the prediction effect of diabetic retinopathy was evaluated and analyzed with the model established with multiple indexes.Results At 50% and 60% of the training data, the improved support vector machine model of particle swarm has achieved accuracy, total accuracy, Matthews correlation coefficient, F1 score, classification accuracy and AUC value in 6 indicators. The optimal results were obtained, and 5 indicators were optimal on 70% of the indicators. The improved model based on genetic algorithm has the best 3 indicators on 80% of the training data.The algorithm based on grid search optimization has achieved 6 optimal results on 90% of the training data.Conclusion The improved support vector machine model has good application value in the prediction and analysis of diabetic retinopathy, and can provide a reference for disease diagnosis.

参考文献/References:

[1]Zhang D,Tang D,Tang L,et al.PCA-SVM-Based Approach of Detecting Low-Rate DoS Attack[C]//International Conference on Computational Intelligence and Software Engineering.2019. [2]Li J,Zhang C,Li Z.Battlefield Target Identification Based on Improved Grid-Search SVM Classifier[C]//International Conference on Computational Intelligence and Software Engineering.2009:1-4. [3]Huang S,Zheng X,Ma L,et al.Quantitative contribution of climate change and human activities to vegetation cover variations based on GA-SVM model[J].Journal of Hydrology,2020(584):124687. [4]Tang H,Wang Z,Wu Y.A multi-fault diagnosis method for piston pump in construction machinery based on information fusion and PSO-SVM[J].Journal of Vibroengineering,2019. [5]董慧妍.基于机器学习的糖尿病性视网膜病变图像分级研究[D].北京邮电大学,2019. [6]Vapnik VN.Controlling the Generalization Ability of Learning Processes[M].The Nature of Statistical Learning Theory.Springer New York,1995. [7]Shenghan Z,Silin Q,Wenbing C,et al.A Novel Bearing Multi-Fault Diagnosis Approach Based on Weighted Permutation Entropy and an Improved SVM Ensemble Classifier[J].Sensors,2018,18(6):1934. [8]Gao Y.Newton method of solving Karush-Kuhn-Tucker systems for a constrained Minimax Problem[J].Journal of Information&Optimization Sciences,2007,28(2):173-184.

相似文献/References:

[1]戴晓敏,丁 勇.分类算法在高血压诊断中的应用研究[J].医学信息,2019,32(01):1.[doi:10.3969/j.issn.1006-1959.2019.01.001]
　DAI Xiao-min,DING Yong.Application of Classification Algorithm in Diagnosis of Hypertension[J].Medical Information,2019,32(02):1.[doi:10.3969/j.issn.1006-1959.2019.01.001]
[2]周靖宇,单慧明,成官迅.基于支持向量机的全乳纹理分析预测乳腺癌新辅助化疗疗效的价值[J].医学信息,2022,35(19):60.[doi:10.3969/j.issn.1006-1959.2022.19.016]
　ZHOU Jing-yu,SHAN Hui-ming,CHENG Guan-xun.Prediction of Neoadjuvant Chemotherapy Outcomes of Breast Cancer Based on Whole Breast Texture with Support Vector Machine[J].Medical Information,2022,35(02):60.[doi:10.3969/j.issn.1006-1959.2022.19.016]
[3]黄辛迪,黄 慧,刘佳俊,等.决策树及支持向量机与深度学习模型在肝癌鉴别诊断中的比较研究[J].医学信息,2023,36(15):70.[doi:10.3969/j.issn.1006-1959.2023.15.012]
　HUANG Xin-di,HUANG Hui,LIU Jia-jun,et al.Comparative Study of Decision Tree, Support Vector Machine and Deep Learning Model in Differential Diagnosis of Liver Cancer[J].Medical Information,2023,36(02):70.[doi:10.3969/j.issn.1006-1959.2023.15.012]

更新日期/Last Update: 1900-01-01