On the streaming video of all typical flow pattern filmings on the experiment system by the high-speed video camera,the information of a single-frame was extracted and made into time-series.The series was analyzed wit...On the streaming video of all typical flow pattern filmings on the experiment system by the high-speed video camera,the information of a single-frame was extracted and made into time-series.The series was analyzed with the non-linear chaotic recurrence plot(RP),less used in the last years.It was combined with the average diagonal length and Shannon entropy of recursive features changed after the increase of gas superficial velocity.The results showed that the information entropy of flow image combined with RP could well characterize the evaluated tract of gas-liquid two-phase flow patterns.At the same time,the information well characterized the volume fraction of all kinds of gas-liquid two-phase flow patterns.The average diagonal length and recursive Shannon entropy of recursive features all increased first and then decreased with the increase of gas superficial velocity,and it reflected the transition of the mechanisms of five typical flow patterns from the recursive characteristics.展开更多
本研究通过特征选择的方法,分析肝癌患者术前临床信息,提高患者的预后模型的准确性。基于多类支持向量机递归特征消除(recursive feature elimination based on multiple support vector machine,MSVM-RFE)方法对进行过肝切除手术的原...本研究通过特征选择的方法,分析肝癌患者术前临床信息,提高患者的预后模型的准确性。基于多类支持向量机递归特征消除(recursive feature elimination based on multiple support vector machine,MSVM-RFE)方法对进行过肝切除手术的原发性肝癌患者的临床变量进行重要特征排序,使用5折交叉验证的支持向量机确定最优特征子集,构造原发性肝癌患者术后的1年、3年无瘤生存和总体生存的列线图。通过与临床医生沟通,确认特征排序结果为合理的。患者3年无瘤生存风险和总生存风险的列线图的一致性指数分别为0.701和0.706。使用多类支持向量机递归特征消除方法后的预测模型准确率有所提高,列线图在临床实践中能够提供患者生存风险信息,简单清晰的反映患者的生存风险。展开更多
文摘On the streaming video of all typical flow pattern filmings on the experiment system by the high-speed video camera,the information of a single-frame was extracted and made into time-series.The series was analyzed with the non-linear chaotic recurrence plot(RP),less used in the last years.It was combined with the average diagonal length and Shannon entropy of recursive features changed after the increase of gas superficial velocity.The results showed that the information entropy of flow image combined with RP could well characterize the evaluated tract of gas-liquid two-phase flow patterns.At the same time,the information well characterized the volume fraction of all kinds of gas-liquid two-phase flow patterns.The average diagonal length and recursive Shannon entropy of recursive features all increased first and then decreased with the increase of gas superficial velocity,and it reflected the transition of the mechanisms of five typical flow patterns from the recursive characteristics.
文摘本研究通过特征选择的方法,分析肝癌患者术前临床信息,提高患者的预后模型的准确性。基于多类支持向量机递归特征消除(recursive feature elimination based on multiple support vector machine,MSVM-RFE)方法对进行过肝切除手术的原发性肝癌患者的临床变量进行重要特征排序,使用5折交叉验证的支持向量机确定最优特征子集,构造原发性肝癌患者术后的1年、3年无瘤生存和总体生存的列线图。通过与临床医生沟通,确认特征排序结果为合理的。患者3年无瘤生存风险和总生存风险的列线图的一致性指数分别为0.701和0.706。使用多类支持向量机递归特征消除方法后的预测模型准确率有所提高,列线图在临床实践中能够提供患者生存风险信息,简单清晰的反映患者的生存风险。