Detecting local convexity on the pareto surface

Recent regain in interest in multi‐criteria optimization approaches to provide a designer with multiple solutions to select from and support decisions has led to various methodologies to generate such solutions and possibly approximate the Pareto set. This paper introduces the notions of H‐ and w‐convexities, and develops a simple method to identify local convexity of Pareto hyper‐surfaces since their shape can dictate the choice of the method to use to obtain Pareto solutions and possibly to build an approximation of that set. The method is based on comparing the results from the weighting method to those of the Tchebycheff method at any point on the Pareto hyper‐surface. If, under some conditions, the points obtained from the two methods are identical or not, a local convexity or its lack can be assumed at that location and in its immediate neighbourhood. A numerical example is included. Copyright © 2002 John Wiley & Sons, Ltd.     

用Delta法估计多维测验合成信度的置信区间

大量研究表明,一般情况下用合成信度可以较好地估计测验信度。对于合成信度及其置信区间的估计方法,在单维测验的情形已有不少研究。但罕有研究讨论多维测验合成信度的区间估计方法。本文用Delta法推导出计算多维测验合成信度的标准误公式,进而计算置信区间,并用一个例子说明如何编程估计多维测验合成信度及其置信区间。     

数学焦虑影响大学生/儿童估算策略运用的对比研究

选取66名大学生和40名小学四年级儿童,运用选择/无选法研究范式,采用自编乘法估算材料考察了数学焦虑对成人和四年级儿童算术策略运用的不同影响.研究结果显示:(1)数学焦虑仅影响成人在自由选择条件中策略选择的正确率.(2)数学焦虑影响儿童的策略分布、策略执行的正确率及最佳选择条件中策略选择的正确率.(3)成人和儿童策略选择的适应性(以有选条件下的策略使用正确率为指标)均受到数学焦虑的影响,低数学焦虑者的适应性明显更好.文章讨论了数学焦虑影响成人与儿童算术策略运用的不同潜在机制.     

练习背景下表征水平的变化

Karmiloff-Smith提出的表征重述理论认为表征重述是人类获取知识的重要途径。为探讨练习是否能够提供表征重述的机会从而促进表征水平的变化,对29名小学一年级儿童进行数字分解组合任务的研究。结果表明：（1）练习背景下约有半数儿童的表征从内隐水平发展到外显水平,但其他儿童的表征水平没有发生任何变化;（2）练习背景下儿童表征变化的路线并不完全像Karmiloff-Smith设想的那样经由程序阶段到元程序阶段再到概念化阶段,而是存在其他多种可能的发展路线;（3）练习背景下儿童获得的表征的灵活性是有限的,即可以推广到近迁移问题上,而难以推广到远迁移问题上     

Asé and Amen,Sister!

At times, the academy seems devoid of justice because it emphasizes the cultivation of knowledge often denied to marginalized individuals and communities. As black queer feminist scholars doing praxis‐driven theorizing from separate fields on the subject of black queer families and communities, we employ research methods that resist the dynamics of power and privilege that exist within normative researcher‐participant exchanges. In this essay, we explore and highlight the ethical, justice‐oriented, and dialogical relationship between researcher‐scholars and research participants. Through story and theory, we illustrate and argue that autoethnographies and narrative interviews can act as epistemological excavation tools for both researchers and participants, as they become sites of individual and collective consciousness. Our work resists capitalist models of research and instead promotes a justice‐oriented and community‐derived building of knowledge.     

SCIENCE AND OTHER COMMON NOUNS: FURTHER IMPLICATIONS OF ANTI-ESSENTIALISM

The term “science” is a common noun that is used to designate a whole range of activities. If Reeves is right—and I think he is—that there is no essence to these activities that allows them to be objectively identified and demarcated from nonscience, then what qualifies as science is determined by communities. It becomes much more difficult on this antiessentialism position to identify and dismiss pseudo-science. I suggest we might find a way forward, though, by engaging a philosophical tradition that has largely been neglected in English-speaking science and religion studies, and by articulating a theory of consensus along the lines of Oreskes (2019).     

An interactive procedure for learning about preferences: Case study of a portfolio manager

This paper presents a new interactive procedure for learning and interpreting the choices available to a decision maker confronted with a discrete decision problem. The procedure guides a decision maker through the evaluation process, has an ability ‘to learn’ from previous choices and generates a set of rules that describe a decision maker's posterior preferences. The validity of the procedure was tested by creating an investment portfolio. The emphasis of the application is on assisting the portfolio manager to determine the decision consistency of a given portfolio selection process.     

一种可融入额外信息的机器学习诊断法

研究将PNN和曼哈顿距离、贝叶斯定理相结合,提出了一种相对简洁的可融入额外信息的认知诊断法MB-PNN,通过模拟和实证研究考察了MB-PNN的有效性和适宜性,得到以下结论:(1)M-PNN的判准率高于PNN,表明将PNN中的ED修改为MD是适宜的;(2)MB-PNN的判准率较M-PNN和PNN高,表明基于多种信息的判别较基于单一信息的判别更为精准;(3)MB-PNN保留了PNN原有的非参数优势,基本不受知识状态分布和样本容量影响;(4)MB-PNN最能区分不同类型的学生,在认知诊断评估实践中更为适宜。     

基于CAT的在线标定:设计与方法

项目增补是题库建设和维护的重要手段, 而标定新题参数是项目增补的重要内容。在线标定设计和在线标定方法分别研究新题的施测方式和参数估计方法, 是计算机化自适应测验(computerized adaptive testing, CAT)情景下项目增补的核心技术。重点厘清在线标定设计与在线标定方法的发展思路和脉络, 并对它们的特点、联系和表现进行介绍和评价。未来应基于其他信息指标进一步研究在线标定设计, 可基于联合估计和误差校正的思路探究在线标定方法, 应加强研究认知诊断CAT和多维CAT的在线标定技术, 深入开展项目增补方法的实证研究。     

‘Liars are less detailed’ …So what? Comparing two recall instructions to detect deception within-subject

Most deception research provides between-subject results (e.g., liars give on average less detailed accounts), which might be of limited value for professionals evaluating credibility on an individual basis. This study examines the optimal instructions of a within-subject multiple recalls strategy to detect deception. A total of 110 participants, divided into a Lie and Truth group, were randomly placed into four interview conditions: two Basic report-everything instructions (1), a Basic recall followed by an Open depth instruction (2), a Basic recall followed by the Verifiability Approach and Information Protocol (3), and two recalls with the Verifiability Approach and Information Protocol (4). All recalls were coded for total details and verifiable details. Group (lie and truth) × Recall (first and second) was only significant in condition 3, with truth tellers providing more verifiable details in the second recall than the first. A simple within-subject decision rule was derived, allowing a 76.9% discrimination rate. Professionals can optimally evaluate credibility using two recalls (Basic recall followed by Verifiability Approach and Information Protocol) and observing the evolution of verifiable details.