科学家绘出世界最大病毒的基因组图谱

科学家绘制出了已知的世界最大病毒的基因组图谱，发现该病毒拥有一些以前被认为只在细菌和其他生物细胞中才存在的基因，这将有助于研究复杂生物的起源。     

岩体可爆性等级判别的随机森林模型及R实现

在引入随机森林分类方法的基础上,选取岩石的容重、抗拉强度、动载强度及岩体完整性系数等指标,建立岩体可爆性等级判别的随机森林模型;在R语言环境下,编写模型数据处理、计算与结果输出的R语言代码,实现可爆性等级判别随机森林模型的计算,得出岩体可爆性等级判定的混淆矩阵,分析各指标对岩体可爆性分级的重要性。研究结果表明:岩体可爆性分级为三、四、六级的判别准确率可达100%;岩石容重对岩体可爆性等级判别的影响最大。这一结果证明了岩体可爆性等级判别的随机森林模型是可靠的,且具有较高的准确率。     

双波段极化雷达遥感图像分类的新方法

充分有效地利用极化雷达遥感数据提取目标散射特征是极化雷达遥感应用研究中的一个重要问题，构造了三个目标概率形式描述：目标相似性参数构造目标平均散射特征描述、目标散射相关阵特征值构造目标散射随机性描述、目标接收总功率构造目标散射强度描述。在目标散射特性的概率形式描述基础上，引入鉴别信息作为目标间的差异量度；这时目标之间的差异可以按照各个波段的平均散射特征、散射随机性和散射强度分别进行量度，为去除冗余信息高效利用数据带来方便。采用最小鉴别信息准则构造了双波段极化雷达遥感图像分类算法及其简化算法；与基于Wishart分布假设的最大似然分类方法相比，无论监督分类还是非监督分类，提出的算法均可以获得更加精确的分类结果。     

基于模型融合的复杂生产过程产品质量预测

复杂生产过程中产品质量与工艺参数间存在复杂的非线性关系,为提高产品质量预测准确性,本文提出了一种基于模型融合的复杂生产过程产品质量预测方法.首先,分别对复杂生产过程建立基于改进随机森林算法的整体预测模型与分段预测模型,其中,针对整体预测模型特征选择问题,提出了一种相关性分析与去冗余处理相结合的特征选择方法,针对分阶段预测模型存在误差累积问题,提出了一种误差修正机制.其次,利用Stacking集成学习算法实现整体预测模型与分阶段预测模型的融合,综合利用二者的预测优势得到对产品质量的最终预测结果.最后,以烟丝生产过程烘丝机入口烟丝含水率的预测为例,通过对比传统单模型的预测方法,验证了本文所提基于模型融合预测方法的准确性.     

基于出口管制清单的质谱仪技术对比分析研究

为研究全球质谱仪制造技术的研发特点，分析我国的竞争优劣势，本文采用监督分类和非监督聚类结合的方法，对全球质谱仪专利技术进行人工智能筛选，得到全球质谱仪制造技术的研究主题布局。将技术研究主题与出口管制清单中所涉及的质谱仪信息进行技术对比，分析中国与美国、欧盟、日本等在质谱仪制造技术方面的研究现状和竞争态势。结果显示随机森林算法对质谱仪制造技术的筛选效果最好。无监督聚类算法将全球质谱仪制造技术聚类为质谱仪整机技术、离子质量分析器技术、质谱仪数据处理模块、进样部件与装置、离子源核心技术等五大模块。通过计量分析发现，近十年来，中国的质谱仪制造技术进步非常快，2015年后，专利已经超过了全球专利数量的一半。中国在进样部件与装置、质谱仪数据处理等研究方向上的技术储备具有明显优势，而在质谱仪整机技术、离子质量分析器技术、离子源核心技术等主题方向上比较薄弱。中国在核心技术方面存在短板，在面对西方技术管制时，存在一定的风险。     

基于SOM神经网络的古代青瓷聚类分析

用EDXRF探针测定了48个来自3个不同产地的4个古代青瓷窑址的瓷片的化学组成．接着用SOM神经网络对胎的主、次量元素化学组成数据进行了聚类分析,结果表明SOM能正确地区分出瓷片的3个产地：龙泉、慈溪和杭州．由于两官窑的瓷片在主、次量元素化学组成上比较接近,使得它们的分类正确率仅为76．92％．于是,作者对两官窑瓷片胎的微量元素化学组成数据用同样的方法又进行了一次聚类分析,结果发现分类正确率提高到了84．61％,说明两官窑的瓷片在微量元素上的差别比在主、次量元素上的差别要大,这与实际情况是一致的．研究表明SOM神经网络方法可以应用于古陶瓷的聚类分析研究中．     

伪随机整数串与伪随机序列产生的新方法

提出了产生伪随机整数串和伪随机序列的一种新方法－全循环小数序列（Ｆ－序列）方法，研究了这类序列的伪随机整数串性质和游程性质，并通过变换得到了一类新的具有最佳自相关函数的序列。     

基于遥感数据的小水库塘坝拦洪计算方法研究与应用

本文以美国陆地资源卫星“LandsatTM／ETM＋”数据为基础,以流域地形分类为手段,以降雨蒸发为控制条件,提出了基于遥感数据的小水库塘坝拦洪计算方法．该方法旨在利用遥感技术研究解决小型水利工程影响流域洪水预报精度的问题,主要针对上游有众多小型水利工程的流域,也可用于无径流资料的小型水库径流估算．以20060826洪水为例,对19个有资料小水库进行拦洪模拟发现,拦洪总量绝对误差为-0．2万m^3,相对误差为-0．12％,模拟精度较好．使用此方法对全流域的小水库塘坝进行拦洪计算,并校正原洪水预报方案,校正后的相对误差由校正前的31．8％降低到10．1％,精度明显提高．     

人类基因组单体型图计划及其"中国卷"

国际人类基因组单体型图计划是针对亚、非、欧裔全基因组中的一百万或更多的单核苷酸序列差异位点进行测定和分析,构建出整合了人类遗传多态信息的每条染色体的“单体型图”,为疾病相关的基因研究提供全球共享的最基本数据。在这一多国参与的重大国际合作项目中,中国将做出10％的贡献,其具体内容为构建3号、21号染色体和8号染色体短臂的单体型图以及提供一半的亚洲样品。中国团队确定的平台技术有基于引物延伸方法的Sequenom的质谱检测系统和PerkinElmer的荧光偏振检测系统以及基于分子杂交原理的Illumina的光纤磁珠检测系统。     

智能文献计量学视角下的知识发现：人工智能+信息计量学

如何高效准确地发现知识并提炼情报是信息管理与图情领域的重要课题。大数据时代人工智能技术的飞速发展为这一课题带来新视野。本文由读《学科信息学与学科知识发现》所悟所感出发,探讨应用于知识发现的文献计量学方法。着眼于文献计量学方法在大规模数据处理、复杂关系发现以及模型可适性、复用性与鲁棒性等问题上的缺陷,本文以智能文献计量学为核心,剖析其与人工智能技术的交互、整合与再创新,并介绍其在生物医学领域知识发现的最新应用。此外,本文还介绍了智能文献计量学在复杂实践中的缺陷,浅谈其发展前景。     

Screening of conformationally constrained random polypeptide libraries displayed on a protein scaffold

The selection of novel proteins or enzymes from random protein libraries has come to be a major objective in current biology, and these enzymes should prove useful in various biological and biomedical fields. New technologies such as in vitro selection of proteins in cell-free systems have high potential to realize evolu tionary molecular engineering of proteins. This review highlights an application of insertional mutagenesis of proteins to evolutionary molecular engineering. Random sequence proteins are inserted into the surface of a host enzyme which serves as a scaffold to display random protein libraries. Constraints on random polypeptide conformations owing to the proximity of N- and C-termini on the scaffold would result in greater screening efficiency of libraries. The scaffold enzyme is also used as a probe for monitoring the hill climbing of random sequence proteins on a fitness landscape and navigating rapid protein folding in the sequence space. Received 9 October 1997; received after revision 6 January 1998; accepted 19 January 1998     

数据流中的近似查询技术

近年来,一种密集型数据的应用得到了广泛的关注。在这种应用中,数据不再来自存储于永久介质的数据库,而是以大量的、连续的、快速的、随时间变化的、无边界的数据流的形式到达。本文深入研究数据流中的近似查询处理问题,针对数据流查询处理的典型特征,平衡内存有限性与查询结果精确性两大关系,综合多个著名系统关于近似查询处理的精髓思想,加以总结提出一系列如窗口查询、直方图、随机采样、小波等近似查询技术。     

无人机图像全自动生成大比例尺真正射影像方法

随着无人机遥感技术逐步从研究开发阶段发展到实际应用阶段,采用无人机图像进行大比例尺的真正射影像生成方法的研究具有重要的现实意义．本文旨在探索将计算机视觉中获得巨大成功的多视三维重建技术应用到对无人机影像处理中,给出了一种基于运动恢复结构重建算法和多视图立体视觉算法全自动生成大比例真正射影像的方法．本文首先分析了无人机图像PMVS重建点云的特点,给出一种由基于面片多视图立体视觉稠密点生成数字表面模型的方法,然后详细介绍了包括正射影像图像坐标映射模型、可见性计算、基于Markov随机场能量优化的面片选择和匀光处理等真正射影像生成的关键步骤．实验结果以及与商业软件的比较表明：本文给出的方法在野外地形和城市区域均能获取有效的真正射影像结果．     

Evaluation of the going-concern status for companies: An ensemble framework-based model

Issuing a going-concern opinion is a difficult and complex task for auditors. The auditors have to take into account different critical factors in order to make the right decision based on information obtained from the auditing process. This study adopts the so-called “random forest” approach (based on the ensemble method) to assist auditors in making such a decision. To investigate the corresponding effect of the proposed approach, we conduct a series of experiments and a performance comparison. The results show that the random forest method outperforms the baseline methods in terms of the accuracy rate, ROC area, kappa value, type II error, precision, and recall rate. The proposed approach is proven to be more accurate and stable than previous methods.     

Randomized unit root processes for modelling and forecasting financial time series: Theory and applications

This paper considers the problems of statistically analysing the levels of financial time series rather than their differences, which are often equivalent to returns and which are traditionally analysed in econometric modelling. This focus on differences is a consequence of the inherent nonstationarity of the levels, and hence analysing the latter requires introducing an alternative framework for modelling nonstationary behaviour. We do this by considering randomized unit root processes, arguing that these can have a natural interpretation in the financial context. The paper thus develops methods for testing for randomized unit roots and for modelling such processes. It then applies these techniques to various financial time series, so as to ascertain their potential usefulness, particularly for forecasting.     

基于复杂网络挖掘的用户行为感知机制

用户行为感知是进行网络管理、安全检测以及应用趋势分析的基础.针对基于流量统计特征检测方法具有计算复杂度高和"概念漂移"的缺陷,提出了一种基于用户复杂网络图的用户行为感知机制算法(UBP-CN).算法将用户标识{IP,Port}和用户交互分别抽象为一个点和一条边,构建了用户复杂网络图;应用社团挖掘算法将复杂网络图划分为互不相交的行为子簇,使得用户之间的通信抽象为一种"社会团体";通过定义基于相对熵的"用户行为模式"(UBM),表征了各个子簇背后表现出的行为特性,并使用"UBM+Port"对各个子簇进行标签映射,实现了用户行为的有效感知.仿真结果表明:在不牺牲用户行为分类准确率的前提下,算法不仅能克服"概念漂移"问题,还能有效降低算法的计算复杂度.     

Genetic studies of diseases

The biological system is a complex physicochemical system consisting of numerous dynamic networks of biochemical reactions and signaling interactions between cellular components. This complexity makes it virtually unanalyzable by traditional methods. Hence, biological networks have been developed as a platform for integrating information from high- to low-throughput experiments for analysis of biological systems. The network analysis approach is vital for successful quantitative modeling of biological systems. The numerous online pathway databases vary widely in coverage and representation of biological processes. An integrated network-based information system for querying, visualization and analysis promised successful integration of data on a large scale. Such integrated systems will greatly facilitate the understanding of biological interactions and experimental verification.     

Studies on the metal-complex of acetyl salicylic acid (aspirin).

The present communication deals with the isolation of acetyl salicylic acid (aspirin complexes with Bi+3 Zn+2 and UO2+2. The characterization of 1:2 complexes have been carried out with the help of conductometric, pH met;ric, elemental analysis and IR spectral studies. Spectrophotometric studies in case of UO2+2 (the only colored complex) in range of 4.2 to 5.5 pH show absorption at 490 nm and complex obey Beers Law at the concentration range of 0.01 M to 0.1 M.     

Expressed protein ligation: a resourceful tool to study protein structure and function

This review outlines the use of expressed protein ligation (EPL) to study protein structure, function and stability. EPL is a chemoselective ligation method that allows the selective ligation of unprotected polypeptides from synthetic and recombinant origin for the production of semi-synthetic protein samples of well-defined and homogeneous chemical composition. This method has been extensively used for the site-specific introduction of biophysical probes, unnatural amino acids, and increasingly complex post-translational modifications. Since it was introduced 10 years ago, EPL applications have grown increasingly more sophisticated in order to address even more complex biological questions. In this review, we highlight how this powerful technology combined with standard biochemical analysis techniques has been used to improve our ability to understand protein structure and function.     

Modelling exchange rate dynamics new perspectives from the frequency domain

In this paper the dynamics of foreign exchange rates is sought to be studied via new frequency domain techniques. Stationarity properties of the rates are analysed via a unit root test as well as a test based on the evolutionary spectrum. Linearity and Gaussianity are analysed via bispectral tests and compared with the more frequently employed time domain tests, such as the McLeod-Li and Tsay tests. Finally, an evaluation of the out-of-sample forecasting properties for eight methods—Random Walk, ARMA, Bilinear, State dependent model, dynamic linear model, ARCH, GARCH, and Garch-in-mean—is made. The methods used here seem to shed a great deal of light on hitherto neglected aspects of exchange rate dynamics.