【摘要】 目的基于智能算法建立卵巢恶性肿瘤标记物组模型,并评价其检验效能。方法以经确诊的143例卵巢恶性肿瘤患者和98例卵巢良性肿瘤患者以及120例健康妇女为研究对象,以CA125、HE4、TSGF、TNF-α、VEGF、CA153、CA199、IL-6 8种血清标记物的水平为自变量,以病理分型结果(恶性肿瘤、良性肿瘤、健康)为因变量,应用支持向量机算法建立卵巢恶性肿瘤标记物组模型,并通过60例测试集评价其效果。结果成功建立了基于支持向量机的卵巢恶性肿瘤标记物组模型,支持向量机模型的诊断准确度为95%。结论支持向量机模型的诊断准确率较高,对卵巢肿瘤的鉴别诊断有重要的参考价值,可做为临床医生的辅助诊断工具。更多还原

【Abstract】 Objective To establish the ovarian cancer marker group model based on intelligent algorithm, to evaluate the power of test. Methods 143 patients diagnosed with ovarian malignant tumors,98 patients diagnosed with benign ovarian tumor,and 120 healthy women were selected as research objects.To setablish ovarian cancer marker group model by the application of support vector machine algorithm,with the independent variables of the serum markers level of 8 kinds of CA125,HE4,TSGF,TNF-α,VEGF,CA153,CA199,IL-6,with the dependent variable by the pathological type results of ovarian malignant tumors,benign ovarian tumor,and healthy,to evaluate the effect through the 60 test suites. Results Ovarian cancer marker group model based on intelligent algorithm has successfully established with the diagnostic accuracy of support vector machine model of 95%. Conclusion Support vector machine model has higher diagnostic accuracy,has important reference value on the differential diagnosis of ovarian neoplasms,is the auxiliary diagnostic tool for clinical doctors.更多还原