一种新的卫星通信网流量预测算法

在通信网络的设计中,使用基于流量预测的网络规划已成为LTE发展的必然趋势。与 地面网络不同,卫星网络由于受资源受限和拓扑时变的不利影响,其流量预测算法必须能兼 顾精度和效率,这令传统的地面网络预测方法已不再适用。为了解决以上问题,提出了一种 新的基于小波回声状态网络的流量预测算法,该算法通过小波多尺度分解的信号处理方法屏 蔽了网络流量的噪声,而后结合了无反馈的回声状态网络联合进行预测。仿真证明,新算法 相比传统算法能大幅提升网络流量的预测精度和运行效率,为卫星网络的流量规划提供了 强有力的决策支持。     

树型网络中IPTV系统的流量均衡策略

提出一种在树型网络中的网络电视(IPTV)系统负载均衡算法,该算法通过服务器放置点的选择和服务器内容分配实现负载均衡.首先把大的树型网络分解成多个小的树型网络,对每个小的树型网络应用一种改进的搜索算法,最后把在多个子树的解组合成原问题的解,较好地将流量均匀分配到整个网络中.     

基于支持向量机的多特征目标抗干扰检测技术

针对箔条干扰时目标与干扰难以区分的问题，设计了一种基于多特征向量的分类算法。该算法首先对目标和箔条的特征进行分析，而后选择并构造了一组具有较高区分度的极化特征识别量，最后采用支持向量机（SVM）方法，通过对特征样本进行训练，获得了较好的分类结果。实验表明，所提算法具有较强的抗箔条干扰能力，且检测正确率可达90%以上。     

基于改进KNN算法的股票风险预测

本文针对投资风险预测问题,测算股票的在险价值并选取指标,采用KNN分类算法对次日股票跌幅进行分类预测,并结合股票不同指标间相关性较强且数值差异过大的特性以及投资者交易实时更新数据的特征,采用向量正交化和更新样本外预测两种方法共同对算法进行改进,在实证对比后发现尽管最终改进前后算法预测准确率均达到85%,但改进后的算法能在K值较小时快速提高预测风险是否发生的正确率,快速捕捉风险,且改进后的算法有效降低了低估风险的概率,在风险控制方面更有效。同时,通过对两种改进方法的分解,发现更新样本外预测这一手段对算法改进的帮助更大。     

联合循环平稳特征PCA与RVM的频谱感知

针对无线信道环境中低信噪比情况下主用户信号检测率较低的问题,提出了一种基于循环平稳特征主成分分析(PCA)与相关向量机(RVM)的认知网络频谱感知算法。该算法结合了主成分分析算法与相关向量机分类方法,应用于解决认知网络频谱感知问题。首先对信号循环平稳特征参数进行特征提取,通过主成分分析进行降维提取信号主成分,生成训练样本和待测样本,并完成对相关向量机的训练,再采用训练完成的相关向量机算法分别对有无主用户情况下的信号进行分类检测,最后获得主用户信号存在性的感知判断。仿真实验表明,与人工神经网络、支持向量机和最大最小特征值算法相比较,所提算法在低信噪比情况下具有较高的分类检测性能,检测率最大可提高61.6%,有效地实现了对主用户信号的感知。     

采用局部二值模式与深度信念网络的人脸识别

针对人脸维度过高和人脸局部特征提取易忽略的问题,提出了一种将多尺度局部二值模式（LBP）算法与深度信念网络(DBN)算法相结合的人脸识别方法。首先采用多尺度LBP算法提取人脸纹理特征,进而将LBP提取的纹理特征作为深度信念网络的输入,最后通过逐层网络训练,得到网络的最优参数,并在ORL人脸库中进行测试,识别率可达95.2%,比使用Gabor小波和主成分分析（PCA）算法的人脸识别高2.6%,说明该算法具有很好的人脸识别能力。     

移动因特网不同流量下切换损失率研究

根据移动因特网中基本的硬切换和改进的半软切换算法时间流程仿真研究了 2种切换在泊松和自相似流量下的切换损失率。仿真数据表明 ,在网络负载均值相同的情况下 ,对于硬切换和半软切换 ,自相似流量的切换损失率要明显低于传统泊松流量 ,并通过分析泊松和自相似流量的概率密度函数尝试给出产生这一区别的原因。     

基于核Fisher鉴别分析的网络入侵检测算法

提出了一种将核Fisher鉴别分析特征抽取与多分类支持向量机算法结合的网络入侵检测技术,扩展了二分类支持向量机．利用经过核Fisher鉴别分析特征抽取后的训练数据构造优化的决策树,从而实现支持向量机的多分类。实验结果表明该算法能够提高检测正确率,同时降低训练时间,取得了良好的效果。     

融合多特征的图像检索算法

针对单一特征不能很好地表述图像的问题,提出了一种融合多特征的图像检索算法。首先,提取查询图像和图像库中样本图像的GIST（Generalized Search Tree)特征,用欧氏距离衡量图像间的GIST相似度值,根据查询图像的GIST特征在图像库中进行检索,将结果按相似度进行排序;然后,提取查询图像和返回结果中前k幅图像的尺度不变特征变换(SIFT)特征,使用BBF(Best Bin First)算法进行特征匹配;最后,通过特征点匹配点对数排序并返回检索结果。实验在改进的Corel1000数据集上进行,与传统的单特征图像检索算法对比,提出的图像检索算法不仅提高了检索准确率,而且获得了较好的检索效率。     

基于层次型Adaboost的动态车牌定位方法

提出了一种结合车牌定位和跟踪的快速动态车牌定位算法。在层次型AdaBoost检测算法的基础上,结合提出的一种基于线性预测的车牌跟踪算法来缩小搜索空间,使得检测速度提高到16帧/秒。还设计了一种结合层次型AdaBoost算法、双阈值判决构造强分类器的检测器训练方法,可以方便地应用于物体定位技术上。     

哈尔滨城市交通发展的问题和对策

过去的20年,哈尔滨获得了前所未有的发展,道路交通建设和交通管理取得了很大成绩。但是,出现的交通拥挤堵塞问题也非常严重。解决城市交通拥堵,应大力发展公共交通,培养一支高素质的城市客运经营管理队伍,建设一系列配套齐全的城市客运服务设施,使哈尔滨的市内交通得以改善。     

Transport safety for older people: A study of their experiences,perceptions and management needs

With evidence of increasing accident risk due to age-related declines in health and cognition affecting driver performance, there is a need for research promoting safe mobility of older people. The present study aimed to identify transport options and licensing issues for a group of older people in an Australian community. Ninety-five participants aged 75 and over were interviewed about their driving status and accident record and tested for cognitive ability. After stratification on cognitive level and driver status (current, ex-driver or non-driver), 30 were selected for further in-depth interviews concerning demographics, licence status and impact of change, travel options available and used, and travel characteristics. Considerable reliance on the motor vehicle as the mode of transport and the decision to cease driving were major quality-of-life issues. There was little evidence of planning and support in making the decision to stop driving. Some differences in transport decisions on the basis of cognitive level were evident; however, people with severely compromised cognitive ability (and, therefore, unable to give informed consent) had been excluded. The study suggested the need for resources to assist older people/carers/health professionals to plan for the transition from driver to non-driver and to manage alternative transport options more effectively.     

Road traffic crashes with fatal and non-fatal injuries in Arkhangelsk,Russia in 2005–2010

The study investigated trends in traffic crashes with fatal and non-fatal injuries in Arkhangelsk, Russia in 2005–2010. Data were obtained from the road police. Negative binomial regression with time regressor was used to investigate trends in monthly incidence rates (IRs) of crashes, fatalities, and non-fatal injuries. During the six-year period, the police registered 4955 crashes with fatal and non-fatal injuries, which resulted in 217 fatalities and 5964 non-fatal injury cases. The IR of crashes with fatal and non-fatal injuries per total population showed no evident change, while the IR per increasing total number of motor vehicles decreased on average by 0.6% per month. Pedestrian crashes constituted 51.8% of studied crashes, and pedestrians constituted 54.6% of fatalities and 44.5% of non-fatal injuries. The IRs of pedestrian crashes and non-fatal pedestrian injuries per total population decreased on average by 0.3% per month, and these were the major trends in the data.     

Reaching those at highest risk to fires and burns: Young children,older adults,and people in low-income communities

Injury specialists have not successfully convinced policy makers and the public that injuries can be controlled. That failure may be due in part to the lack of a unified understanding of injury control. The two most important models utilized in injury control are Haddon's Matrix1 and the Public Health Approach (PHA). This paper argues that the PHA should be combined with the two axes of Haddon's Matrix to result in a model that is coherent and comprehensive. Thus it is better than either one of the original models on their own. Haddon's Matrix has two axes. The first includes elements of the epidemiological triad, host, vector, and environment and likens injury to disease. The second axis includes three time intervals, pre-event, event, and post-event. The importance of including time was that injury was conceptualized as predictable and preventable. The weakness of Haddon's matrix is that it lacks a systematic plan of action. The Public Health Approach is a methodology for addressing injury, which consists of a hierarchy of four levels; surveillance, risk factor identification, intervention evaluation and program implementation. The use of the PHA with no specific orientation or means of application is its weakness. The PHA lacks a systematic point of application. Haddon's Matrix lacks a systematic action plan. Therefore we propose the PHA as the systematic strategy for the more theoretical framework of Haddon's matrix. By combining these concepts a coherent and comprehensive three-dimensional framework is defined. The unified model closes the potential gaps in the two original models and includes a systematic approach not previously achieved. This unified model is practical in defining individual studies and groups of studies. It can be used as an inventory, for a complete understanding of a particular injury. Diagrams of the model are presented to help teach the concepts of injury described in this unitary model. In conclusion, we can say that the inclusion of three injury concepts in one framework provides a rigorous and coherent construct for the understanding of injury and implementation of control activities. It can therefore be used to design more comprehensive programs for injury control and promote policies and funding commensurate with the magnitude of the injury problem.     

The key risk indicators of road traffic crashes in Serbia,Niš region

Traffic crashes are a serious public health and socio-economic problem, and they are one of leading causes of death of males in Serbia. In order to determine the main causes of road traffic crashes in Ni? region, Republic of Serbia, nine key risk indicators of road traffic crashes are identified, organized into four groups (number and structure of crashes, consequences of traffic crashes, temporal and spatial location of traffic crashes, causes and sanctions), and described in detail. The indicators are ranked based on group fuzzy analytic hierarchy process (AHP), where experts from insurance companies, professionally engaged in the process of risk assessment, with different previous experience present different impact on the final ranking. In relation to the most influential factors in the assessment of traffic crashes, improper and unadjusted vehicle speed, overtaking and passing, as well as the physical condition of the driver have the highest effect.     

Road crash injuries and fatalities in Isfahan,Iran from March 2006 to March 2009

With rapid development of social economies, road traffic accidents have continued to increase, and have become the primary public hazard to humans. The main goal of the present study was to investigate road traffic crash (RTC) fatalities and injuries in the city of Isfahan, Iran. A sample of 150,940 accident cases was considered from Isfahan Police Safety Driving Department, involving drivers and passengers of all ages, and covering a 3-year period. The record linkage identified 24,608 drivers and passengers injured or died as a result of RTC in the city of Isfahan over the 3-year period. The finding of this study shows that the highest rate of RTC fatality was 40% and 58%, which comprises the male drivers and female passengers within the age classes 25–34 and 35–44, respectively. On average, there were one death every 3 days and every hour, someone was injured and taken to an emergency department for RTCs in the city of Isfahan. The highest men to women death and injured ratios were 4:1 and 2:1, respectively. The use of seat belt devices in our population was worrisome. The article ends with a number of recommended measures for the improvement of road safety.     

Relative effects of store traffic and customer traffic flow on shopper spending

This article examines the relative effects of store traffic and customer traffic flow on shopper spending in a single study. An analysis of Croatian hypermarket survey data indicates that store traffic alone does not adequately explain shopper spending. Instead, gross customer traffic flow and realized customer traffic flow are stronger drivers of money spent. The article contributes to the retailing literature by clarifying the respective roles of store traffic and customer traffic flow. Recognizing the greater revenue-generating effect of in-store traffic flow, retailers should better design their store layout and merchandizing strategies to optimize traffic movement and boost store performance.     

1xEV-DO系统数据流量研究

本文首先分析了 1xEV -DO前向业务信道的链路特性和今后将面临的新业务特性 ,然后着重对 1xEV -DO的无线资源分配和平均业务流量作了研究。     

Understanding road traffic risks from the street hawker’s perspective

This study explored street hawkers’ lived experiences and risks in traffic using the phenomenological case study design. Data from 24 hawkers at three hawking spots and field observations form the basis of our study. A stratified purposeful sampling scheme was employed to ensure a balance of gender and age. Thematic data analytic strategy was employed in the data analysis. We found that participants’ traffic incidents were mainly self-induced, attributable to their negligence on the road, indiscriminate road crossing and car-following behaviour. Our results suggest a perceived relationship between socio-demographic characteristics and personal traffic accident encounters. Also, we noted that the ability to manage traffic risk seem to influence hawkers’ traffic risk perceptions, with gender and age as possible mediators. As this study is exploratory, we recommend further scientific enquiry to explain these issues and to provide the basis for appropriate interventions to manage the road safety implications of this activity.