Multicategory classification via discrete support vector machines

Discrete support vector machines (DSVM), originally proposed for binary classification problems, have been shown to outperform other competing approaches on well-known benchmark datasets. Here we address their extension to multicategory classification, by developing three different methods. Two of them are based respectively on one-against-all and round-robin classification schemes, in which a number of binary discrimination problems are solved by means of a variant of DSVM. The third method directly addresses the multicategory classification task, by building a decision tree in which an optimal split to separate classes is derived at each node by a new extended formulation of DSVM. Computational tests on publicly available datasets are then conducted to compare the three multicategory classifiers based on DSVM with other methods, indicating that the proposed techniques achieve significantly higher accuracies. This research was partially supported by PRIN grant 2004132117.     

高速粒子场同轴Fraunhofer全息数据处理系统研究

介绍了用同轴Fraunhofer全息测量强动载下材料表面微喷射粒子场的数据处理方法.在全息像重建过程中，将再现的三维粒子场由计算机控制分成许多小薄层采集.在数据处理过程中根据图像中粒子边缘，决定在整幅图或局部采用边缘检测的方法提取大粒子，根据图像的灰度分布将图像分成很多小区域，在每个小区域采用不同的阈值分割图像.在处理结果的校正中根据粒子场的特点，去除过大、过小和重复统计粒子.采用该方法得到了粒子的空间分布图像及粒子大小的统计结果.     

On constant elasticity of substitution – Constant elasticity of transformation Directional Distance Functions

This paper aims to present non-linear CES (Constant Elasticity of Substitution)–CET (Constant Elasticity of Transformation) Directional Distance Functions. These measures inherit the structure of the standard Directional Distance Functions and that of the CES–CET technology. These functions allow non-parametric estimation of efficiency scores through linear programming method. Besides, the CES –CET technology gives the opportunity to explore α-returns to scale assumption for the new distance functions. The duality theory is investigated through pseudo profit, cost and revenue functions. The dual standpoint provides non-linear adjusted prices that can occur into non-linear pricing practices. An application is proposed to give an illustrative example of the primal CES–CET Directional Distance Functions.     

Dynamic discovery of query path on the lattice of cuboids using hierarchical data granularity and storage hierarchy

Analytical processing on multi-dimensional data is performed over data warehouse. This, in general, is presented in the form of cuboids. The central theme of the data warehouse is represented in the form of fact table. A fact table is built from the related dimension tables. The cuboid that corresponds to the fact table is called base cuboid. All possible combination of the cuboids could be generated from base cuboid using successive roll-up operations and this corresponds to a lattice structure. Some of the dimensions may have a concept hierarchy in terms of multiple granularities of data. This means a dimension is represented in more than one abstract form. Typically, neither all the cuboids nor all the concept hierarchy are required for a specific business processing. These cuboids are resided in different layers of memory hierarchy like cache memory, primary memory, secondary memory, etc. This research work dynamically finds the most cost effective path from the lattice structure of cuboids based on concept hierarchy to minimize the query access time. The knowledge of location of cuboids at different memory elements is used for the purpose.     

双波段光谱融合的猪肉多品质参数同时检测方法研究

在双波段可见/近红外光谱系统(350~1 100和1 000~2 500 nm)中,由于两台仪器性能有所不同,导致在波段重叠区域对同一样品测得的反射率不同,出现数据交叉现象。针对此问题,提出一种波段连接数据融合的方法,以期对两个波段的光谱进行更好的应用。首先采集60个生鲜猪肉样品表面的反射光谱信息,利用Savitzky-Golay(S-G)平滑和标准正态变量变换进行预处理,然后利用单一波段和双波段光谱数据与猪肉品质参数(颜色参数L*, a*, b*,pH和蒸煮损失率)理化值建立偏最小二乘预测模型,并分析比较。利用提出的波段融合方法对两个波段重叠区域出现的交叉进行处理,处理后的双波段光谱融合数据对参数L*, a*, b*, pH以及蒸煮损失率建模,验证集的相关系数分别为0.948 8,0.920 0,0.950 5,0.930 1和0.903 5,模型效果与未融合前相当甚至更优。采用无信息变量消除法方法进行特征变量筛选,利用优选后的特征变量建立了更为简化的模型。实验结果表明,所提出的波段融合方法能够对两个波段光谱数据实现较好的融合,利用融合后的光谱数据有利于建立更简化、性能更佳的预测模型。     

A General Framework for Mining Massive Data Streams

In many domains, data now arrive faster than we are able to mine it. To avoid wasting these data, we must switch from the traditional “one-shot” data mining approach to systems that are able to mine continuous, high-volume, open-ended data streams as they arrive. In this article we identify some desiderata for such systems, and outline our framework for realizing them. A key property of our approach is that it minimizes the time required to build a model on a stream while guaranteeing (as long as the data are iid) that the model learned is effectively indistinguishable from the one that would be obtained using infinite data. Using this framework, we have successfully adapted several learning algorithms to massive data streams, including decision tree induction, Bayesian network learning, k-means clustering, and the EM algorithm for mixtures of Gaussians. These algorithms are able to process on the order of billions of examples per day using off-the-shelf hardware. Building on this, we are currently developing software primitives for scaling arbitrary learning algorithms to massive data streams with minimal effort.     

数字微镜哈达玛光谱仪谱线弯曲的分析与修正

由于数字微镜（digital micro-mirror device, DMD)哈达玛变换光谱仪其成本低,光能利用率高及无运动部件等优势,逐渐成为光谱仪领域的研究重点。研制了一款基于DMD的哈达玛变换光谱仪。为了解决光谱仪谱线弯曲造成的光谱分辨率下降的问题,对基于DMD的哈达玛变换光谱仪中的谱线弯曲所引起的谱带混叠进行了分析。首先,导出了谱带混叠与谱线弯曲的关系式。然后,提出了两个过程来解决谱带混叠,一是通过调整DMD编码条纹,使DMD所编码的谱带最大限度地与标准谱带重合; 二是通过数据处理对谱带混叠进行修正。最后,通过对谱线曲率半径为5.8×104,7.8×104和9.7×104 μm等六种情况下谱带混叠进行了分析与修正,拟合出光谱混叠和修正效果与谱线曲率半径的关系。结果表明： 对于不同程度的谱线弯曲经过这两个过程修正后,分辨率都会改善到接近光学系统的分辨率, 说明这两个过程对修正谱线弯曲具有普适性、并且方法简单、有效。     

Efficiency and sustainability assessment for a group of farmers in the Brazilian Amazon

The aim of this paper is to use DEA models to evaluate sustainability in agriculture. Several variables are taken into account and the resulting efficiency is measured by comparison. The performance of family farms is analysed here (variables: farmed area, work force, and production). As agricultural sustainability depends on the maintenance of systems of production for long periods of time, the models were run for the years of 1986 and 2002. Tiered DEA models were used to group farmers in sustainability categories. Non-parametric regression models were used to identify the factors affecting the efficiency measurements. All the results indicate that the majority of the farmers increased their efficiency along the time. These improvements may support the existence of sustainability.     

核实验数据获取系统中的触发信号预处理

简要介绍了在核物理实验数据获取系统中,用FPGA芯片和VHDL语言实现对事件触发信号的预处理。阐述了在核物理实验数据获取系统中对事件触发信号的一般要求; 结合60Co伽马谱测量实验,介绍了基于一套自主研发的数据获取系统中触发信号的预处理,并给出了用VHDL语言实现该触发信号预处理的思想、仿真时序图和实验结果。该方法通过实验验证,证明切实可行,灵活性强。     

Fluid phase topology of ethanol+benzene+cyclohexane at 101.3 kPa

In this article, isobaric vapor–liquid equilibria for the ternary mixture of ethanol?+?benzene?+?cyclohexane was experimentally investigated at atmospheric pressure. Vapor–liquid equilibria data for ethanol?+?benzene?+?cyclohexane at 101.3?kPa were obtained with a Othmer-type ebulliometer. Data were tested and considered thermodynamically consistent. The experimental results showed that this ternary mixture is completely miscible and exhibits three binary homogeneous azeotropes and a ternary minimum azeotrope at the studied conditions. Satisfactory results were obtained for correlation of equilibrium compositions with UNIQUAC activity coefficients model and also for prediction with UNIFAC method. In both cases, low root mean square deviations of vapor mole fraction and temperature were calculated. The capability of ethanol as modified distillation agent at atmospheric condition is discussed in terms of the thermodynamic topological analysis. However, owing to the complex topology of the ternary mixture it leads to a distillation scheme with three columns and difficult operation and thus, ethanol is not recommended as a separating agent for benzene?+?cyclohexane azeotrope.