论文部分内容阅读
目的:建立用于筛选有效处方的疗效评价模型。方法:相似度计算方法采用Jaccard相似度和Ochiai相似度两种算法,采用ROC曲线进行模型参数估计。结果:共入选54例慢性消化道疾病门诊患者72次复诊病历,相邻两次诊疗的Jaccard相似度>0.5(ROC曲线下面积为0.815),Ochiai相似度>0.65为相似,(ROC曲线下面积为0.861)。结论:在临床疗效评价中可以采用Ochiai相似度疗效评价模型,临床相邻两次处方的Ochiai相似度>0.65时可认为前一次诊疗有效。
Objective: To establish a curative effect evaluation model for screening effective prescriptions. Methods: The similarity calculation method uses Jaccard similarity and Ochiai similarity, and uses ROC curve to estimate the model parameters. Results: There were 72 referral cases in 54 outpatients with chronic digestive diseases, Jaccard similarity between adjacent two treatments> 0.5 (area under ROC curve was 0.815), Ochiai similarity> 0.65 were similar (area under ROC curve 0.861). Conclusion: Ochiai similarity evaluation model can be used in the evaluation of clinical curative effect. The Ochiai similarity of 0.65 between two adjacent clinical prescriptions can be regarded as effective for the previous diagnosis and treatment.