中日古典园林哲学比较

中日两国古典园林的哲学基础都是儒、道、佛三家，但三家在中日两国园林中表现强弱有差异。中国突出儒家思想，以至于园林显出入世治世和诗情画意的文人气质；日本突出佛家思想，显出弃世厌世和枯禅清淡的僧徒气质。但是，两国园林都崇尚道家的自然山水观。     

Obtaining the consensus and inconsistency among a set of assertions on a qualitative attribute

It is well understood how to compute the average or centroid of a set of numeric values, as well as their variance. In this way we handle inconsistent measurements of the same property. We wish to solve the analogous problem on qualitative data: How to compute the “average” or consensus of a set of affirmations on a non-numeric fact, as reported for instance by different Web sites? What is the most likely truth among a set of inconsistent assertions about the same attribute?Given a set (a bag, in fact) of statements about a qualitative feature, this paper provides a method, based in the theory of confusion, to assess the most plausible value or “consensus” value. It is the most likely value to be true, given the information available. We also compute the inconsistency of the bag, which measures how far apart the testimonies in the bag are. All observers are equally credible, so differences arise from perception errors, due to the limited accuracy of the individual findings (the limited information extracted by the examination method from the observed reality).Our approach differs from classical logic, which considers a set of assertions to be either consistent (True, or 1) or inconsistent (False, or 0), and it does not use Fuzzy Logic.     

Bias associated with the discriminant analysis approach to the estimation of mixing proportions

The estimation of the mixing proportion π₁, in which the first of two multivariate normal groups occur is considered on the basis of a sample drawn from a mixture of them. Under the model studied, there are also data of known origin available from each of the groups for the formation of a discriminant rule R. The bias of the estimator of π₁, based on the proportion of the mixture data assigned to the first group by R, is investigated. Also, the case where the data of known origin are taken to be correlated is considered.     

基于模式间混淆关系自适应构建层次化分类器

为了解决层次化分类器的设计难点——子分类器的层属关系及其内部组成的确定,本文首先定义了模式间混淆关系,用于描述不同模式在判决域中的相互作用;进而提出了基于混淆矩阵度量模式间混淆关系的方法。设计并实现了多模式混淆关系分析机MPCRAM,将有师指派和无师自组两种常用的模式重组方法有机结合,遵循Fisher准则,自适应地产生层次化分类器结构。大量综合测试证实了该方法有效、实用,可显著提高分类器的识别性能和稳健性。     

A novel algorithm to segment foreground from a similarly colored background

Color similarity between the background and the foreground causes most moving object detection algorithms to fail. This paper proposes a novel algorithm designed to segment the foreground from a similarly colored background. Central to this algorithm is that the motion cue of the moving object is useful for foreground modeling. We predict the position of the moving object in the current frame using historical motion information, and then use the prediction information to construct a predictive model. The mixture foreground model is a union of the predictive model and the general foreground model. Final segmentation is obtained by combining a likelihood modification technique and the mixture foreground model. Experimental results on typical sequences show that the proposed algorithm is efficient.     

基于K-L散度的最大后验弧主导的混淆网络生成算法

为快速生成高质量混淆网络,该文提出一种最大后验弧主导的快速生成算法。它只需遍历一遍Lattice,具有线性时间复杂度。采用K-L散度(Kullback-Leibler Divergence,KLD)来度量弧标号之间的发音相似性,改善了混淆网络生成中弧对齐的准确性。实验结果显示,所提算法在生成速度上和Xue的快速算法是可比的,而生成质量更好。通过采用KLD作为弧标号相似性测度,生成混淆网络的质量得到了进一步提高。     

面向乳腺癌图像的浅层高识别卷积神经网络研究

乳腺癌是易发生且致死率高的恶性肿瘤之一,及早诊断识别是降低致死率的关键.基于应用广泛的乳腺癌病理图像,结合卷积神经网络展开乳腺癌的识别研究.针对癌症图像细节和纹理特征难以识别的问题,采用插值处理将图像进行适当放大,以便研究分析.针对卷积神经网络参数庞大不易训练和不易硬件实现的问题,提出一种精简的5卷积层W型网络结构,具...     

Matching performance of vehicle icons in graphical and textual formats

The current research classified 82 vehicle icons into seven categories (image-related, concept-related, semi-abstract, arbitrary, abbreviation, word, and combined) for their matching accuracy, matching sequence, and matching time. These data can be compared and used as a framework for future icon development. Forty participants, all with a university degree, took part in this experiment. Half of the participants had intensive driving experience, while the other half never driven a car. The results indicated that on average, word icons had a significantly greater matching accuracy than the other icon formats; ranging from 4.7 to 20.8% difference. Regarding the matching sequence, participants matched image-related icons before other icon formats. Arbitrary and combined icons took significantly longer to match than other icon formats by 1.4–6.2 s. Based on the high matching accuracy (86.3%) and high ratings on subjective design features, word format can be used for functions describable using simple English for users with English reading ability. Confusion matrices showed that 63.2% of the misunderstandings were caused by similarity in format or function.     

Efficacy of bi-component cocrystals and simple binary eutectics screening using heat of mixing estimated under super cooled conditions

The values of excess heat characterizing sets of 493 simple binary eutectic mixtures and 965 cocrystals were estimated under super cooled liquid condition. The application of a confusion matrix as a predictive analytical tool was applied for distinguishing between the two subsets. Among seven considered levels of computations the BP-TZVPD-FINE approach was found to be the most precise in terms of the lowest percentage of misclassified positive cases. Also much less computationally demanding AM1 and PM7 semiempirical quantum chemistry methods are likewise worth considering for estimation of the heat of mixing values. Despite intrinsic limitations of the approach of modeling miscibility in the solid state, based on components affinities in liquids under super cooled conditions, it is possible to define adequate criterions for classification of coformers pairs as simple binary eutectics or cocrystals. The predicted precision has been found as 12.8% what is quite accepted, bearing in mind simplicity of the approach. However, tuning theoretical screening to such precision implies the exclusion of many positive cases and this wastage exceeds 31% of cocrystals classified as false negatives.     

Detecting of the extremely low frequency magnetic field ranges for laptop in normal operating condition or under stress

This paper presents a new method for analysis and management of data, which are obtained by measuring the magnetic field emission in the laptop neighborhood. Accordingly, the extremely low frequency magnetic field radiation emitted from laptop is considered. Laptops operate in normal condition or under stress condition. Measurement is performed on top and bottom measuring positions of laptops, fed by alternating current and battery. Results are analyzed and discussed based on standard safe limit. Then, magnetic field ranges are obtained from clustering of measuring positions by K-Medians algorithm for detection of high emission levels. A comparison between ranges in normal condition and ranges under stress condition is pursued by the confusion matrix, revealing that higher magnetic field ranges are obtained when laptop is under stress. Finally, investigation of the confusion matrix confirms the distribution of measuring positions from lower range to higher range from normal to stress condition.