问卷设计知识管理中的范例推理

将范例推理技术应用于问卷设计知识重用过程中,提出了基于范例推理的问卷设计知识重用系统的应用框架,研究了问卷设计CBR系统中的一些关键技术:范例的表示、范例组织、范例检索算法、范例修正优化和重用、范例学习等.范例检索采用KNN法与遗传禁忌算法相结合的混合检索机制,通过建立GA适应度函数模型以及选择、禁忌交叉、禁忌变异算子等自动确定权重.我们组织和建立了初步的范例知识库,进行了相关实验.结果表明,融合遗传算法的范例检索方法可较大程度上缩短设计时间、提高问卷设计效率和质量,并有效地支持问卷设计的知识重用和知识管理创新.     

Correcting grade deflation caused by multiple-choice scoring

In a study involving three sections of 181 pre-calculus students at a four-year college, partial credit scoring on multiple-choice questions was examined over an entire semester. Questions were constructed by experienced teachers to pose carefully designed, incorrect alternatives, many of which implicitly suggested varying degrees of partial knowledge or understanding. Computing students' course grades based solely on percentage correct scoring resulted in a course-wide GPA of 1.38. This seemed inconsistent with the fact that students were screened for entry into the course. To avert apparent grade deflation due to such right/wrong scoring, student scores were recomputed based on supplementing percentage correct scores with partial credit. This resulted in a course-wide increase in GPA to 1.92. Using that same letter grade distribution, and curving percentage correct averages to achieve the 1.92 GPA, it was found that 10% of the students were given different grades from those determined by the partial credit scoring. Regarding the efficacy of awarding partial credit points, it was found that partial credit by itself measured achievement of relevant mathematical skills that were missed by percentage correct scoring. Overall, grades determined by partial credit scoring seemed more reflective of both the quantity and quality of student knowledge than grades determined by either traditional percentage correct scoring or any corrective curving.     

A Hessian Regularized Nonlinear Time Series Model

We introduce a nonparametric nonlinear time series model. The novel idea is to fit a model via penalization, where the penalty term is an unbiased estimator of the integrated Hessian of the underlying function. The underlying model assumption is very general: it has Hessian almost everywhere in its domain. Numerical experiments demonstrate that our model has better predictive power: if the underlying model complies with an existing parametric/semiparametric form (e.g., a threshold autoregressive model (TAR), an additive autoregressive model (AAR), or a functional coefficient autoregressive model (FAR)), our model performs comparably; if the underlying model does not comply with any preexisting form, our model outperforms in nearly all simulations. We name our model a Hessian regularized nonlinear model for time series (HRM). We conjecture on theoretical properties and use simulations to verify. Our method can be viewed as a way to generalize splines to high dimensions (when the number of variates is more than three), under which an analogous analytical derivation cannot work due to the curse of dimensionality. Supplemental materials are provided, and will help readers reproduce all results in the article.     

Computational singular perturbation with non-parametric tabulation of slow manifolds for time integration of stiff chemical kinetics

This paper presents a novel tabulation strategy for the adaptive numerical integration of chemical kinetics using the computational singular perturbation (CSP) method. The strategy stores and reuses CSP quantities required to filter out fast dissipative processes, resulting in a non-stiff chemical source term. In particular, non-parametric regression on low-dimensional slow invariant manifolds (SIMs) in the chemical state space is used to approximate the CSP vectors spanning the fast chemical subspace and the associated fast chemical time-scales. The relevant manifold and its dimension varies depending on the local number of exhausted modes at every location in the chemical state space. Multiple manifolds are therefore tabulated, corresponding to different numbers of exhausted modes (dimensions) and associated radical species. Non-parametric representations are inherently adaptive, and rely on efficient approximate-nearest-neighbor queries. As the CSP information is only a function of the non-radical species in the system and has relatively small gradients in the chemical state space, tabulation occurs in a lower-dimensional state space and at a relatively coarse level, thereby improving scalability to larger chemical mechanisms. The approach is demonstrated on the simulation of homogeneous constant pressure H₂–air and CH₄–air ignition, over a range of initial conditions. For CH₄–air, results are shown that outperform direct implicit integration of the stiff chemical kinetics while maintaining good accuracy.     

Asymptotic properties for median cross-validated nearest neighbor median estimate in nonparametric regression

Consider the nonparametric regression model , whereg is an unknown function to be estimated on [0, 1], are the fixed design points in the interval [0, 1] and is a triangular array of row iid random variables having median zero. The nearest neighbor median estimator is taken as the estimator of the unknown functiong(x). Median cross validation (mev) criterion is employed to select the smoothing parameterh. Leth _π ^* be the smoothing parameter chosen by mev criterion. Under mild regularity conditions, the upper and lower bounds ofh _π ^* , the rate of convergence and the weak consistency of the median cross-validated estimate are obtained. Project supported by the National Natural Science Foundation of China and the Doctoral Foundation of Education of China.     

Test sets of integer programs

This article is a survey about recent developments in the area of test sets of families of linear integer programs. Test sets are finite subsets of the integer lattice that allow to improve any given feasible non-optimal point of an integer program by one element in the set. There are various possible ways of defining test sets depending on the view that one takes: theGraver test set is naturally derived from a study of the integral vectors in cones; theScarf test set (neighbors of the origin) is strongly connected to the study of lattice point free convex bodies; the so-calledreduced Gröbner basis of an integer program is obtained from a study of generators of polynomial ideals. This explains why the study of test sets connects various branches of mathematics. We introduce in this paper these three kinds of test sets and discuss relations between them. We also illustrate on various examples such as the minimum cost flow problem, the knapsack problem and the matroid optimization problem how these test sets may be interpreted combinatorially. From the viewpoint of integer programming a major interest in test sets is their relation to the augmentation problem. This is discussed here in detail. In particular, we derive a complexity result of the augmentation problem, we discuss an algorithm for solving the augmentation problem by computing the Graver test set and show that, in the special case of an integer knapsack problem with 3 coefficients, the augmentation problem can be solved in polynomial time.Supported by a Gerhard-Hess-Forschungsförderpreis of the German Science Foundation (DFG).     

Efficient calculation of configurational entropy from molecular simulations by combining the mutual-information expansion and nearest-neighbor methods

Changes in the configurational entropies of molecules make important contributions to the free energies of reaction for processes such as protein-folding, noncovalent association, and conformational change. However, obtaining entropy from molecular simulations represents a long-standing computational challenge. Here, two recently introduced approaches, the nearest-neighbor (NN) method and the mutual-information expansion (MIE), are combined to furnish an efficient and accurate method of extracting the configurational entropy from a molecular simulation to a given order of correlations among the internal degrees of freedom. The resulting method takes advantage of the strengths of each approach. The NN method is entirely nonparametric (i.e., it makes no assumptions about the underlying probability distribution), its estimates are asymptotically unbiased and consistent, and it makes optimum use of a limited number of available data samples. The MIE, a systematic expansion of entropy in mutual information terms of increasing order, provides a well-characterized approximation for lowering the dimensionality of the numerical problem of calculating the entropy of a high-dimensional system. The combination of these two methods enables obtaining well-converged estimations of the configurational entropy that capture many-body correlations of higher order than is possible with the simple histogramming that was used in the MIE method originally. The combined method is tested here on two simple systems: an idealized system represented by an analytical distribution of six circular variables, where the full joint entropy and all the MIE terms are exactly known, and the R,S stereoisomer of tartaric acid, a molecule with seven internal-rotation degrees of freedom for which the full entropy of internal rotation has been already estimated by the NN method. For these two systems, all the expansion terms of the full MIE of the entropy are estimated by the NN method and, for comparison, the MIE approximations up to third order are also estimated by simple histogramming. The results indicate that the truncation of the MIE at the two-body level can be an accurate, computationally nondemanding approximation to the configurational entropy of anharmonic internal degrees of freedom. If needed, higher-order correlations can be estimated reliably by the NN method without excessive demands on the molecular-simulation sample size and computing time.     

结合多尺度空间滤波和层级网络的基于结构保持的高光谱特征选择

为了充分利用高光谱图像蕴含的丰富的光谱信息和空间信息,提出了结合多尺度空间滤波和层级网络的基于结构保持的高光谱特征选择算法.算法利用基于l2,1范数的数学模型,选出同时保存全局相似性结构和局部流形结构的特征子集;在多个尺度的窗口中使用双边滤波,自适应计算滤波核,自动在光谱数据中融入空间信息,增强了类内相似性和类间相异性,避免了参量选择;引入层级结构实现空间信息和光谱信息的深入融合,提高了分类准确度;讨论了层级数目和窗口尺度个数对分类准确度的影响.在Indian Pines和PaviaU两个数据集的实验表明,该算法在大部分地物种类上的分类准确度都有较大幅度的提升,总体分类准确度分别达到90.98%和94.20%,相比其他方法明显提高了地物分类准确度.     

Pattern recognition of Chinese flue-cured tobaccos by an improved and simplified K-nearest neighbors classification algorithm on near infrared spectra

In tobacco industry of China, tobacco leaves are classified and managed in terms of their cultivation areas and plant parts of tobacco-stalks. However, sometimes intentionally or involuntary mislabeling cultivation areas, blending tobacco plant parts would occur into tobacco market. The error will affect the style and quality of cigarettes. In the present work, more than 1000 Chinese flue-cured tobacco leaf samples, which have 12 genotypes and cultivated from 5 to 10 regions of China in 2003 and 2004, have been discriminated by means of an improved and simplified KNN classification algorithm (IS-KNN) based on near infrared (NIR) spectra. An original method of optimizing number of significant principal components (PCs) based on analysis of error and cross-validation was advanced. Compared with conventional pattern recognition methods KNN, NN, LDA and PLS-DA, IS-KNN exhibits good adaptability in discrimination of complicated Chinese flue-cured tobaccos. The practice in this work shows that optimized number of PCs and performance of classification models are closely relative to complicated extent of samples but not to number of categories or samples. The results demonstrated the usefulness of NIR spectra combined with chemometrics as an objective and rapid method for the authentication and identification of tobacco leaves or other kinds of powder samples.     

Coalescence of Cluster Beam Generated Sub‐2 nm Bare Au Nanoparticles and Analysis of Au Film Growth Parameters

In this work is presented the growth model for Au films grown on a carbon substrate at room temperature by using as building blocks Au nanoparticles (NPs) with 1.4 nm mean size generated via remote cluster beam synthesis and soft landing on the substrate. The key results highlighted in this work are that 1) the deposited nanoparticles coalesce at substrate level in such a way that the film growth is 3D, 2) newly formed nanoparticles at substrate level are predominantly magic number clusters and 3) coalescensce takes place as soon as two neighboring nanopartciles come closer than a critical distance. The film growth was investigated by TEM as a function of Au load, in the range 0–1.2 μg/cm². Two distinct regimes are identified: the “landing regime” and the “coalescence regime”. During the latter the film growth is 3D with a dynamic scaling exponent z of 2.13. Particular attention was devoted to the study of the evolution of the NP population from the moment they are generated with the cluster beam generator to the moment they land on the substrate and coalesce with other NPs. Our results show that 1) the NPs generated by the cluster beam are heterogeneous in size and are made by more than 95% by Au Magic numbers, mainly Au₂₀ and Au₅₅ and 2) kinetic processes (coalescence) at substrate level is capable of producing NPs populations made of larger Au magic numbers containing up to several thousands of Au atoms. Experimental and simulation results provide insight into the coalescence mechanism and provide strong evidence that the NPs coalesce when the nearest neighbor distance is below a critical mark. The critical distance is at its minimum 0.4‐0.5 nm and it is still unclear whether it is constant or not although the best matching simulation results seem to point to a superlinear dependence from the NP size difference between two neighboring candidate coalescing NPs. The coalescence phenomenon investigated in this work pinpoints the unique self‐organization properties of these small Au NPs in creating films with a stable edge‐to‐edge mean nearest neighbor distance of the order of 1.4 nm.