基于GA_XGBoost模型的大学生科研能力预测问题研究

科学评价大学生科研创新能力对我国科研水平的提高具有重要意义.采用机器学习模型来预测大学生科研能力可以起到良好的效果,提出一种GAXGBoost模型来实现对大学生的科研能力预测.此模型是以Xgboost算法为基础,然后充分利用遗传算法的全局搜索能力自动搜索Xgboost最优超参数,避免了人为经验调参不准确的缺陷,最后采用精英选择策略以此确保每一轮都是最佳的进化结果.通过分析表明,所采用的GAXGBoost模型在大学生科研能力预测的结果中具有很高的精度,将此模型与Logistic Regression、Random Forest、SVM等模型进行对比,GAXGBoost模型的预测精度最高.     

Managing differences by focusing on communication qualities: Pupils learning mathematics in pairs at a computer

The article presents a study of face-to-face verbal communication when pupils work with geometry in pairs and share a stand-alone computer with Geoboard software. The work sessions of pairs of 9th grade pupils are video recorded, and a screen recorder is used to capture their computer activity. The study focuses on identifying and characterizing good communication practices that enhance mathematics learning through joint research reflections between the pupils, the teacher, and the researcher. The major finding concerns how two of the pupils have diverging perspectives on how to communicate mathematics, yet they manage to have productive collaboration. The differences generate collaborative challenges for the pupils and an opportunity to discuss how pupils’ communication qualities can prove important in fruitfully managing such challenges.     

英国国家科研与创新署学科交叉研究资助机制及启示

Newly established in 2018, the UK Research and Innovation (UKRI) strengthens the strategic coordination of the UK research and innovation system by bringing together seven Research Councils, Research England, and Innovate UK. Through its nine organizations, UKRI funds multidisciplinary and interdisciplinary research in a number of priority areas. It also runs the Strategic Priorities Fund to support multidisciplinary and interdisciplinary research in strategic areas identified by government policies as well as the Global Challenges Research Fund to promote challenge-led interdisciplinary research needed by developing countries. The UKRI makes significant efforts to engage stakeholders in the development, design, and implementation of multidisciplinary and interdisciplinary programs. It has also developed a range of mechanisms to improve the evaluation of multidisciplinary and interdisciplinary projects. Chinese science and innovation funding agencies could draw upon the UKRI experience from four aspects to advance interdisciplinary research in China.     

有机化学实验线上课程之扬长避短教学设计与实践篇

It has been two months since a boom of online education triggered by the epidemic in China. At present, we are keeping focus on how to optimize our online class. In the case of chemistry laboratory courses, there's not much that can be done to experimental operations through online teaching. While for the traditional teaching procedure, there is still room for improvement in terms of integrating research to teaching, interactivity, etc. This paper will present some design strategies for improving teaching the organic chemistry laboratory online. To be specific, it describes how teaching materials like the lesson plan and virtual lab were coordinated into the online teaching. And we will also discuss the holistic approach to a better outcome for students' active learning and integration research into teaching by redesigning multiple phases, such as the pre-laboratory preparation, live online class, experimental operations.     

面向远海岛礁群的双向物流网络规划

针对若干远海岛礁群系统封闭性的分布特点,将中心岛礁选址、各岛礁建设泊位数量和规模、仓库储存量、运输船型和数量以及航线设置、班期组织等作为规划内容,以整个运输网络系统总成本最低为原则,综合考虑了仓库存储理论,建立了兼顾“选址-库存-路径”的规划模型。结合所建模型特点,提出了双层并行搜索遗传算法,构建了以海上运输为基本交通方式的双向物流系统。最后,以我国南海为例进行建模并求解,采用不同算法并通过多组不同规模的算例进行演算比较,验证了所建模型和算法的有效性与合理性。     

科研成果向教学实验的转化设计与实践成效*

围绕学校高水平应用型人才培养目标,依托省级重点实验室与省级化学化工大学生创新训练中心等科研教学平台,将“零排放型化学镀镍及其废液资源化为掺杂型LiFePO₄/C复合材料”的科研成果进行拆解和提炼,设计为分阶段、多层次的10个教学实验项目,并予以实践。结果表明,该科教融合模式营造出良好的科研氛围,有利于学生科研素养与创新意识的培养、实践能力与创新能力的提升。     

Measurement of preferences with self-explicated approaches: A classification and merge of trade-off- and non-trade-off-based evaluation types

Self-explicated approaches are popular preference measurement approaches for products with many attributes. This article classifies previous self-explicated approaches according to their evaluation types, i.e. trade-off- versus non-trade-off-based, and outlines their advantages and disadvantages. In addition, it proposes a new method, the presorted adaptive self-explicated approach that is based on Netzer and Srinivasan’s (2011) adaptive self-explicated approach and that combines trade-off- and non-trade-off-based evaluation types. Two empirical studies compare this new method with the most popular existing self-explicated approaches, including the adaptive self-explicated approach and paired comparison preference measurement. The new method overcomes the insufficient discrimination between importance weights, as usually found in non-trade-off-based evaluation types; discourages respondents’ simplification strategies, as are frequently encountered in trade-off evaluation types; is easy to implement; and yields high predictive validity compared with other popular self-explicated approaches.     

Mathematical modelling and the learning trajectory: tools to support the teaching of linear algebra

In this article we present a didactic proposal for teaching linear algebra based on two compatible theoretical models: emergent models and mathematical modelling. This proposal begins with a problematic situation related to the creation and use of secure passwords, which leads students toward the construction of the concepts of spanning set and span. The objective is to evaluate this didactic proposal by determining the level of match between the hypothetical learning trajectory (HLT) designed in this study with the actual learning trajectory in the second experimental cycle of an investigation design-based research more extensive. The results show a high level of match between the trajectories in more than half of the conjectures, which gives evidence that the HLT has supported, in many cases, the achievement of the learning objective, and that additionally mathematical modelling contributes to the construction of these linear algebra concepts.     

指纹物质中毒品检测的研究进展

指纹作为案件现场中最常见的物证之一,不仅包含了可用于人身识别的形态特征,还蕴含了丰富的反映遗留人人身特点的化学信息。通过对指纹中毒品物质的检测及分析,可提供指纹遗留者毒品的吸食、接触情况,将其与特定案件相联系,反映个人生活方式信息,为涉毒人员的发现提供新思路。该文利用文献计量学方法总结了该领域的发展脉络以及研究现状与热点,主要总结了目前已建立的指纹物质中毒品分析检测的各种技术方法和特点,并对未来研究方向及难点进行评述,以期为实际案件指纹中的毒品检测提供参考。     

Five Years of ChemRxiv: Where We Are and Where We Go From Here**

ChemRxiv was launched on August 15, 2017 to provide researchers in chemistry and related fields a home for the immediate sharing of their latest research. In the past five years, ChemRxiv has grown into the premier preprint server for the chemical sciences, with a global audience and a wide array of scholarly content that helps advance science more rapidly. On the service's fifth anniversary, we would like to reflect on the past five years and take a look at what is next for ChemRxiv.