基于非线性偏最小二乘方法的急性电离辐射损伤早期伤情分类研究

发布时间：2018-07-31 06:57

【摘要】：在电离辐射事故发生后，对大量辐射暴露人群进行迅速、准确地伤情分类对于伤员的早期治疗和事故的定级都非常重要。本文采用基于核函数的非线性偏最小二乘方法，从剂量效应和时间效应两个方面，考察了急性电离辐射对大鼠血浆中氨基酸代谢的影响，，并通过遗传算法筛选出若干辐射损伤估计指标。结果表明，辐射后5h，对无辐射暴露的伤情分类正确率达到92.3%，筛选出6个辐射损伤估计指标：Arg, Phe, Gln, Thr, Ile, Cit；辐射后24h，不同程度辐射损伤的伤情分类正确率都在84%以上，筛选出5个辐射损伤估计指标：Tyr, Gln, Hyp, Thr,Leu；辐射后72h的分类效果最佳，不同程度辐射损伤的伤情分类正确率均达到了90%以上，筛选出8个辐射损伤估计指标：Trp, Cy2, Phe, Asn, Gly, Lys, Cit,Orn。研究表明，基于核函数的非线性偏最小二乘方法相较线性偏最小二乘方法模型解释能力和预测能力均有较大提高，更适合于分析关系复杂的大鼠血浆氨基酸数据，有助于提高伤情分类的速度和准确度。
[Abstract]:After the accident of ionizing radiation, it is very important for the early treatment of the wounded and the classification of the accident to classify the injury situation quickly and accurately for a large number of people exposed to radiation. The effects of acute ionizing radiation on the metabolism of amino acids in plasma of rats were investigated in terms of dose effect and time effect by using nonlinear partial least square method based on kernel function. Some indexes of radiation damage estimation were screened by genetic algorithm. The results showed that the correct rate of classification of injuries without radiation exposure reached 92.3% at 5 h after radiation, and the correct classification rate of different degrees of radiation damage was more than 84% by screening out six radiation damage indexes: 1: Arg, 24 h after Phe, Gln, Thr, Ile, Cit; radiation. Five radiation damage estimation indexes: 1) Tyr, Gln, Hyp, ThrrLeu, 72h after radiation, the classification accuracy rate of different degrees of radiation damage were above 90%, and 8 radiation damage estimation indexes: TRP, Cy2, Phe, Asn, Gly, Lys, Citry Orn. were screened out. The results show that the nonlinear partial least squares method based on kernel function is better than the linear partial least squares method in interpretation and prediction, and is more suitable for analyzing the plasma amino acid data of rats with complex relationship. It helps to improve the speed and accuracy of injury classification.
【学位授予单位】：苏州大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R818;O212.1

【参考文献】