非法疫苗风险区的空间范围预测

发布时间：2018-05-19 11:13

本文选题：非法疫苗 + MAXENT模型　；参考：《山东科技大学》2017年硕士论文

【摘要】：自“山东济南非法经营疫苗系列案件”披露后,该事件影响范围广泛,引发社会各界人士的高度关注和重视,引起公众、受种者和儿童家长对涉案疫苗安全性和有效性的担忧。科学严谨地分析涉案疫苗接种在中国可能波及的空间范围,为采取后续处置措施提供依据,保障受种者的健康,尽快恢复公众接种疫苗的信心,是本研究的主要目的。在生态学中,解决物种分布适宜区的问题通常采用物种分布模型(species distribution models,SDMS)推测物种在某一地区的适生性分布。本次研究借鉴物种适宜性分布预测方法,分别利用最大熵模型(MAXENT)和规则集遗传算法模型(GARP)在地理空间上对非法疫苗在中国地区的潜在分布范围进行了预测。MAXENT模型是根据物种现实分布点和物种分布区域的环境预测物种的潜在分布区,其结果要优于同类的其他预测模型。本文应用生态位模型和GIS技术对涉案疫苗在中国区域内可能波及范围进行估计,明确涉案疫苗风险区的空间分布及主要影响因子,以期为政府采取后续处置措施提供依据。并利用受试者工作特征曲线比较不同模型模拟精度。实验结果表明:1 )MAXENT模型和GARP模型都较好的预测了非法疫苗的风险区,山东、河北、河南、江苏、安徽是主要的受涉案疫苗影响的高风险区。2)人口密度、产业结构是影响非法疫苗风险区空间分布的主要环境因子。3)GARP模型得到的结果分析:产生了连续范围较大的潜在分布区,在没有非法疫苗的地区也产生了过多预测的破碎化分布;而MAXENT预测到的潜在分布区,在不同区域具有不同的环境适生性指数,而且成功地排除了不合理的破碎化分布,与采集到的真实数据对比分布区吻合度较高,从而更准确地展示了非法疫苗流入的潜在分布格局。
[Abstract]:Since the disclosure of the case of "the illegal Operation of Vaccine Series in Jinan, Shandong Province", the incident has affected a wide range of people from all walks of life, and has aroused public concern and concern about the safety and effectiveness of the vaccine involved, as well as among the public, grantees and parents of children. The main purpose of this study is to scientifically and rigorously analyze the space range of the vaccination involved in China, to provide the basis for the subsequent disposal measures, to protect the health of the seed recipients and to restore public confidence in the vaccination as soon as possible. In ecology, species distribution models are usually used to predict the suitable distribution of species in a region. This study draws lessons from the prediction method of species suitability distribution, The maximum entropy model (MAXENT) and the rule set genetic algorithm (GARP) model are used to predict the potential distribution of illegal vaccines in China in geographical space. The MAXENT model is based on the species distribution point and species distribution area. Potential distribution of environmentally predicted species, The results are superior to other prediction models of the same kind. In this paper, niche model and GIS technique are used to estimate the potential spread of the vaccine in China, and to determine the spatial distribution and main influencing factors of the vaccine risk area, so as to provide the basis for the government to take further measures to dispose the vaccine. The simulation accuracy of different models was compared by using the operating characteristic curve of the subjects. The experimental results show that both the 1: 1 Maxent model and the GARP model are good predictors of the population density in the risk areas of illegal vaccines, Shandong, Hebei, Henan, Jiangsu, and Anhui are the main high-risk areas affected by the vaccines involved. Industrial structure is the main environmental factor affecting the spatial distribution of illegal vaccine risk areas. The results obtained from the GARP model are as follows: a large range of potential distribution areas have been generated, and too many predicted fragmentation distributions have been produced in areas without illegal vaccines; The potential distribution area predicted by MAXENT has different environmental suitability index in different regions, and the unreasonable fragmentation distribution is successfully excluded, which is consistent with the real data collected. Thus a more accurate picture of the illegal vaccine inflow potential distribution pattern.
【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R95;P208

【参考文献】