一种基于AIS和Bayes网络的垃圾邮件过滤算法

将人工免疫系统和贝叶斯网络相结合应用于垃圾邮件的过滤,设计了一个基于AIS和Bayes网络的垃圾邮件过滤算法,分析并解决了该算法实现过程中的关键问题,给出了亲和力的计算方法和抗原/抗体的结构定义。在PU1bare 语料的基础上与Carreras提出的AdaBoost方法进行对比实验,最后给出了实验结果。     

基于动态人工免疫的邮件分类算法研究

使用虚拟基因库技术,对用于邮件分类的人工免疫系统（AISEC）进行改进,提出了动态人工免疫分类算法（DAICA）模型,改进了AISEC的抗体更新过程。当分类正确时,充分利用参与正确分类的抗体,快速改善抗体质量;当分类错误时,不再是简单地将参与错误分类的抗体直接移去,而是对这些移去的抗体进行体细胞高频变异,以保持先前遇到的抗原信息。还研究了新算法DAICA中使用的参数α与β对算法性能的影响。实验表明,这种改进可以提高分类准确率。     

基于T细胞免疫耐受的检测器生成算法设计

通过模拟T细胞的培育机制,包括阴性选择和阳性选择,提出基于T细胞免疫耐受的检测器生成算法.该算法采用基于规则的基因表达方式同匹配规则相结合的方法,能够灵活表示分类规则间的或关系,并且引入阴性选择算子,使算法能够更有效地生成检测器.     

响应面法优化猕猴桃果浆酶解工艺参数研究

本文研究果胶酶反应温度和作用时间及酶用量对猕猴桃果浆出汁率、果汁澄清度和乙醇不溶物的影响,采用通用旋转设计和响应面分析猕猴桃果浆酶解最优条件。结果表明,猕猴桃果浆的最佳酶解条件为:反应温度为45℃,作用时间为113min,果胶酶用量为0.013%,在此条件下,猕猴桃果浆预测出汁率达到83.30%,乙醇不溶物0.98g/100ml,果汁的澄清度为81.47%。验证试验表明,响应面优化后得到猕猴桃果浆出汁率、果汁澄清度和乙醇不溶物回归方程的相关系数分别为0.958、0.904和0.922。     

新一代GPM IMERG卫星遥感降水数据在中国南方地区的精度及水文效用评估

新一代GPM(Global Precipitation Measurement)IMERG(Integrated Multi-satellit E Retrievals for GPM)卫星遥感反演降水数据产品提供了覆盖范围更大、精度和时空分辨率更高的新的降水数据来源。本文以北江流域为例,基于高精度融合降水数据产品定量评估了新一代准实时"early-run"和"late-run"IMERG产品(IMERG-E和IMERGL)以及非实时后处理的"final-run"IMERG产品(IMERG-F)的精度,并结合VIC(Variable Infiltration Capacity)分布式水文模型评估了该系列产品的水文效用。结果表明:(1)研究区域内非实时IMERG-F产品有着较高的精度,网格尺度上日尺度相关系数为0.65,相对偏差为5.87%,优于上一代3B42-V7产品,而准实时的IMERG-E及IMERG-L产品的日相关系数则为0.6左右,精度同样令人满意;流域平均尺度上各产品精度均进一步提高;(2)情景I(基于站点观测降雨数据率定模型)的水文模拟结果表明,非实时IMERG-F产品表现出较好的水文效用,纳西效率系数为0.622;准实时的IMERG-E及IMERG-L产品虽总体上表现较差,但在汛期表现较好,说明该产品有应用于短期洪水预报的潜力;(3)情景II(基于3B42-V7数据率定模型)的结果表明,各IMERG产品的水文模拟表现均有显著改善,说明3B42历史数据适合在使用IMERG产品进行水文模拟时用于水文模型率定。     

Artificial immune system for optimal design of composite hydrogen storage vessel

Based on the clonal selection principle in the biological immunology, an artificial immune system (AIS) is proposed to optimize the weight of Al-carbon fiber/epoxy composite hydrogen storage vessels under burst pressure constraint. The AIS simulates the generation of the antigens, the combination of the antigens and antibodies, and the selection, cloning and hypermutation of B cells, the maintenance of diverse antibody cells, the generation of the memory cells with high affinities and the death of the antibody cells with low affinities. Effects of the antibody size and the iterative number on the optimization results are explored. By comparison, the AIS shows higher search velocity and precision than the genetic algorithm and simulated annealing.     

AIS系统时隙预约选择算法的分析

STDMA是AIS（船舶自动识别系统）中的关键技术，它采用时隙动态预约的自组织方式工作，由于候选时隙的重叠，系统可能发生时隙的冲突，针对这一问题，利用概率分析理论，推导出在两船及多船的情况下，系统时隙冲突概率的数学模型，给出了仿具的曲线并进行了分析，结果表明STDMA时隙动态预约算法采用周期时隙预约和增量时隙预约相结合的方式，减少了时隙发生冲突的问题，最后，提出了对这一算法进行改进的思路。     

Crash costs by body part injured, fracture involvement, and threat-to-life severity. United States, 2000

This paper presents costs per US motor vehicle crash victim differentiated into many more diagnostic categories than prior estimates. These unit costs, which include the first keyed to the 1990 edition of Abbreviated Injury Scale (AIS) threat-to-life severity scores, are reported by body part, whether a fracture/dislocation was involved, and the maximum AIS score among the victim's injuries. This level of detail allows for a more accurate estimation of the social costs of motor vehicle crashes. It also allows for reliable analyses of interventions targeting narrow ranges of injuries. The paper updates the medical care data underlying the US crash costs from 1979 to 1986 to the mid 1990s and improves on prior productivity cost estimates. In addition to presenting the latest generation of crash victim costs, this paper analyzes the effects of applying injury costs classified by AIS code from the 1985 edition to injury incidence data coded with the 1990 edition of AIS. This long-standing practice results in inaccurate cost-benefit analyses that typically overestimate benefits. This problem is more acute when old published costs adjusted for inflation are used rather than the recent costs.     

AIS网络数据链路容量分析及拥塞解决方法

研究AIS网络的容量与拥塞控制,对于维护AIS网络的稳定、设计完善的AIS通信终端有重要的意义;分析了AIS(Automtical Identification System)网络的链路容量,推算了AIS网络信道中通信终端的最大容量,以及该极限容量的变化规律;论述了当AIS网络处于数据链负荷程度危及安全信息发射时的拥塞解决方法,给出了拥塞控制规则的流程图,有效地解决了AIS网络在拥塞时的冲突问题;该算法已用于笔者设计的AIS通信设备,实验表明,该算法能有效维护AIS网络在拥塞状态下的顺利运行,可用于各种AIS设备中.     

Priorities of pedestrian protection—A real-life study of severe injuries and car sources

The aim of this study was to aid the optimisation of future, vehicle based, pedestrian injury countermeasures. The German In-Depth Accident Study (GIDAS) database was queried for pedestrians impacted by the front of a passenger car or van. A total of 1030 cases from 1998 to 2008 were studied including 161 severely (AIS3+) injured pedestrians. Considering the severe injuries, the most frequent injury mechanisms were “leg-to-front end”, “head-to-windscreen area”, “chest-to-bonnet area”, and “chest-to-windscreen area”. For children, a “head-to-bonnet area” impact was the second most common source of injury. With safety systems targeting these five injury mechanisms, 73% (95% confidence interval [CI], 65–81%) of the severely injured pedestrians would be provided protection from all of their vehicle-induced severe injuries. Omitting the windscreen area, this figure is decreased to 44% (CI, 36–53%). Furthermore, 31% of the surviving pedestrians were estimated to sustain a permanent medical impairment at any level. For more severe impairment, head was the dominating body region. The study shows that when developing countermeasures for the windscreen area to mitigate head injuries, attention should be paid to the structural parts of the windscreen area with a special focus on brain injuries. Finally, the incidence and risk of severe injury were derived as functions of impact speed for different body regions and injury sources.