基于支持向量机的类风湿性关节炎诊断模型

Rheumatoid arthritis diagnostic model based on support vector machine algorithm

目的:建立类风湿性关节炎诊断的数学模型。方法:采用支持向量机的最优二分类方法,构建类风湿性关节炎诊断的支持向量机模型,对100例类风湿性关节炎患者和50例非类风湿性关节炎其他风湿免疫病患者的类风湿因子和抗环瓜氨酸肽抗体2项指标进行检验和仿真诊断。结果:jackknife检验法和5维交叉验证法检测样本的诊断正确率均为86%。结论:可以用支持向量机建立类风湿性关节炎的诊断系统。

Objective To establish a mathematical model for the diagnosis of rheumatoid arthritis. Methods A rheumatoid arthritis diagnostic model based on support vector machine optimal classifier was established. The values of rheumatoid factor and anti-cyclic cirtrullinated peptide of 100 rheumatoid arthritis patients and 50 controls were trained and simulated. Results The correct diagnostic rate of the samples obtained by the jackknife test and 5-fold cross validation was 86~. Conclusion The classifier based on support vector machine algorithm can be used to establish rheumatoid arthritis diagnostic system.