The Use of Resampling Methods to Simplify Regression Models in Medical Statistics

The number of variables in a regression model is often too large and a more parsimonious model may be preferred. Selection strategies (e.g. all-subset selection with various penalties for model complexity, or stepwise procedures) are widely used, but there are few analytical results about their properties. The problems of replication stability, model complexity, selection bias and an over-optimistic estimate of the predictive value of a model are discussed together with several proposals based on resampling methods. The methods are applied to data from a case–control study on atopic dermatitis and a clinical trial to compare two chemotherapy regimes by using a logistic regression and a Cox model. A recent proposal to use shrinkage factors to reduce the bias of parameter estimates caused by model building is extended to parameterwise shrinkage factors and is discussed as a further possibility to illustrate problems of models which are too complex. The results from the resampling approaches favour greater simplicity of the final regression model.     

Bell不等式实验检验的重大进展——Weihs实验和Rowe实验述评

Weihs实验（1998）和Rowe实验（2001）可能是自Aspect实验（1982）以来已完成的几十个Bell型实验中最重要的两个。这两个实验已分别地封闭了定域性漏洞和效率漏洞,但定域隐变量理论并没有完全被排除。相信有朝一日人们会完成一个无漏洞的实验。即使如此,定域隐变量理被排除的问题仍然会是悬而未决的。断言试图通过完成一个无争议的判决性实验来解决定域隐变量理论与量子力学的冲突是不可能的。     

基于Windows的主机入侵检测系统联动防火墙的设计与实现

设计与开发了一个运行在windows200下的轻量级入侵检测系统,系统采用VC＋＋6.0作为开发工具,利用WinPcap技术来捕获网络数据包。在数据包处理方面采用了统计与特征相结合的检测方法,从整体结构而言是一个基于主机的入侵检测系统。文章重点介绍系统总体结构和联动防火墙模块的设计和实现,最后以攻击检测实验加以验证。     

Deceptive Detection Methods for Effective Security with Inadequate Budgets: The Testing Power Index

Detection of contraband depends on countermeasures, some of which involve examining cargo containers and/or their associated documents. Document screening is the least expensive, physical methods, such as gamma ray detection are more expensive, and definitive manual unpacking is most expensive. We cannot apply the full array of methods to all incoming cargoes, for budgetary reasons. We study the problem using principles of game theory, and find that best detection rates are achieved when the available budget is allocated between screening and definitive unpacking using a mixture of strategies that maximize detection rate and, further, serve to deceive opponents as to the specific tests to which contraband will be subjected. This yields increases of as much as 100% in detection, with essentially no increase in inspection cost.     

线性预测盲源提取算法应用于阵列雷达的分析

通过深入分析阵列雷达的杂波/干扰抑制问题,提出应用基于线性预测的有噪盲源提取算法来提取阵列雷达的回波信号,实现杂波/干扰的同时抑制,最终提高阵列雷达的目标探测能力。在建立逼真的阵列雷达接收信号模型的基础上,通过仿真分析了该算法提取阵列雷达回波的能力和改善阵列雷达的目标探测能力。结果表明,该算法提取雷达回波信号的效果理想,平均相关系数大于0.93;当运动目标的归一化多谱勒频率大于±0.1Hz时,阵列雷达的检测概率接近1,验证了该算法的可行性和有效性。     

信号检测论在内隐学习研究中的应用

信号检测论引入心理学研究领域是对传统心理物理学方法的一个发展.在内隐学习实验中,信号检测论的两个指标d'和β分离了个体的感受性和主观判断标准,利用d'和β的共变关系证实了内隐学习的存在和内隐学习的特征.本文探析了内隐学习研究中应用信号检测论(正态分布信号检测模型)的逻辑机理.从被试的心理行为反应上微观地分析了用SDT揭示无意识领域的可行性.     

Marginal Trust in Risk Managers: Building and Losing Trust Following Decisions Under Uncertainty

Although a considerable amount of research has examined correlates of baseline public trust in risk managers, much less research has looked at marginal changes in public trust following specific events. Such research is important for identifying what kinds of events will lead to increases and decreases in public trust and thus for understanding how trust is built and lost. Using a taxonomy based upon signal detection theory (SDT), the current article presents two experimental studies examining marginal trust change following eight different types of events. Supporting predictions, cautious decisionmakers who accepted signs of danger (Hits and False Alarms) were more likely to be trusted than those who rejected them (All Clears and Misses). Moreover, transparency about an event was associated with higher levels of marginal trust than a lack of transparency in line with earlier findings. Contrary to predictions, however, trust was less affected by whether the decisions were correct (i.e., Hits and All Clears) or incorrect (i.e., False Alarms and Misses). This finding was primarily due to a "False Alarm Effect" whereby Open False Alarms led to positive increases in trust despite being incorrect assessments of risk. Results are explained in terms of a cue diagnosticity account of impression formation and suggest that a taxonomy of event types based on SDT may be useful in furthering our understanding of how public trust in risk managers is gained and lost.     

格语法在大学英语教学中的语义贡献

“格语法”理论是美国著名语言学家菲尔墨在1968年首次提出的。格语法理论旨在考察句法结构和语义释义之间的关系。尽管格语法理论的研究在句法领域尚未取得丰硕成果，但对语义学的研究和发展却做出了卓越的贡献。它在英语教学中得到广泛应用，在帮助师生区分歧义句、纠错等方面发挥着重要作用。