【摘要】 目的将Logitboost算法用于判别分析慢性乙型肝炎肝脏纤维化血清学检测指标,建立一种无创性慢性乙型肝炎纤维化分期诊断模型。方法 226例慢性乙型肝炎患者进行肝活检和肝纤维化血清学指标的检测,血清学指标进行相关系数和Logitboost算法判别分析,并与肝活检结果比对。结果 Logitboost算法在迭代100轮时判别正确率最高,与肝活检的结果相比较,纤维化分期S(0～4)的诊断符合率分别为92.3%、89.5%、91.5%、96.9%、97.7%。结论 Logitboost算法可以对慢性肝炎肝纤维化进行有效分期。更多还原

【Abstract】 Objective To use Logitboost algorithm for discriminant analysis of serological detection indexes in hepatic fibrosis of chronic hepatitis B and to establish a non-invasive staging diagnosis model in hepatic fibrosis of chronic hepatitis B.Methods 226 patients with chronic hepatitis B were tested for liver biopsy and liver fibrosis serological indexes detection.The discriminant analysis of serological indexes was performed on the correlation coefficients and Logitboost algorithm and compared with the resultsof liver biopsy.Results Logitboost algorithm provided most accuracy in iterative 100 rounds,and compared with the results of liver biopsy,the diagnostic coincidence rates in the stages of fibrosis S(0-4) were 92.3%,89.5%,91.5%,96.9% and 97.7% respectively.Conclusion Logitboost discriminant can stage liver fibrosis of chronic hepatitis effectively.更多还原