一种基于SVM的产品质量鉴别方法

为解决现实生产中某些产品无法用客观分析方法进行准确品质鉴别的问题,弥补人工感观鉴定方法的不足,提出了基于SVM的产品质量鉴别方法,通过产品的化学成份完成品质分类鉴定。实验采用葡萄酒的数据做仿真计算,结果表明,该方法可以对葡萄酒的品质进行有效的鉴定,准确度较高。     

动态利率期限模型的估计和实证研究

利率期限结构的估计在金融研究中有着重要的地位，它是资产定价．金融产品设计、保值和风险管理的基准。该文第一次选择货币市场基准利率R007为样本数据，对一系列单因子利率模型进行欧拉离散、运用极大似然估计法进行了参数估计，并对参数估计结果进行了统计检验和模型间的似然比检验。其中，对于由最大似然估计法必须求解的复杂似然方程组，采用了不同的最优化方法进行了有益的尝试，发现蒙特卡洛法特别适合求解这类问题。最后，发现CKLS模型是最适合于描述我国短期回购利率的动态变化的，R007的变化表现出明显的均值回复特征。     

数据挖掘技术在产品相似性上的应用研究

市场细分导致的产品多样化使预测与计划更加困难,而相似性产品的确定可以改善库存结构,促进销售的增长。文章在论述数据库设计和基础数据准备的基础上,提出了三种方法用于确定相似性产品,并给出了实例,最后作了比较分析。     

Evaluation of MODIS SWIR and NIR-SWIR atmospheric correction algorithms using SeaBASS data

Using the NASA maintained ocean optical and biological in situ data that were collected during 2002-2005, we have evaluated the performance of atmospheric correction algorithms for the ocean color products from the Moderate Resolution Imaging Spectroradiometer (MODIS) on Aqua. Specifically, algorithms using the MODIS shortwave infrared (SWIR) bands and an approach using the near-infrared (NIR) and SWIR combined method are evaluated, compared to the match-up results from the NASA standard algorithm (using the NIR bands). The in situ data for the match-up analyses were collected mostly from non-turbid ocean waters. It is critical to assess and understand the algorithm performance for deriving MODIS ocean color products, providing science and user communities with the important data quality information. Results show that, although the SWIR method for data processing has generally reduced the bias errors, the noise errors are increased due mainly to significantly lower sensor signal-noise ratio (SNR) values for the MODIS SWIR bands, as well as the increased uncertainties using the SWIR method for the atmospheric correction. This has further demonstrated that future ocean color satellite sensors will require significantly improved sensor SNR performance for the SWIR bands. The NIR-SWIR combined method, for which the non-turbid and turbid ocean waters are processed using the NIR and SWIR method, respectively, has been shown to produce improved ocean color products.     

Water and heat regimes modelling for a vast territory using remote-sensing data

Methods have been developed to assess water and heat regime characteristics of a vast agricultural region for vegetation season based on the model of land-surface–atmosphere interaction. The model is adjusted to assimilate estimates of the land-surface and meteorological characteristics derived from Advanced Very High Resolution Radiometer/National Oceanic and Atmospheric Administration (AVHRR/NOAA), Moderate Resolution Imaging Spectroradiometer/Earth Observing System Terra & Aqua (MODIS/EOS Terra and Aqua), and Scanning Enhanced Visible and Infrared Imager/Meteosat-9 (SEVIRI/Meteosat-9) data. The case study has been carried out for the agricultural Central Black Earth region of European Russia of 227,300 km² for the 2009–2012 vegetation seasons. The methods of satellite data processing have been developed or refined, which provide the retrieval of vegetation characteristics, land-surface temperature, and precipitation. The techniques for the assimilation of satellite-based products in the model have been developed. Some major water regime characteristics have been generated such as soil water content, evapotranspiration, and others.     

Land-use mapping and change detection in a coal mining area - a case study in the Jharia coalfield,India

In an area like the Jharia coalfield (JCF), where extensive and rapid underground and opencast mining is going on continuously, land-use studies are of paramount importance. This paper discusses the remote sensing-GIS techniques used for identification of various land-use classes on satellite imagery and enhanced products and identification of time-sequential changes in land-use patterns. The various land-use classes, recognised from satellite image data and field surveys, are dense vegetation, sparse vegetation, fire, opencast mining (coal), overburden dump, subsidence and barren wasteland, settlement, transport network, river and water pond. A number of image processing operations have been carried out on remote sensing data for enhancing land-use patterns. It has been found that Landsat TM false colour composites (FCC) of bands 4, 3 and 2; FCC of bands 7, 5 and 3; FCC of bands 5, 4 and 2 and ratio images provide very useful information for land-use mapping. The normalised difference vegetation index (NDVI) images have been used for vegetation studies. Image characters of various land-use classes on black-and-white and enhanced colour products have been tabulated. Land-use maps of selected windows have been prepared and examples given. Time-sequential surface changes that have occurred in the JCF since 1975 and particularly between November 1990 to November 1994 have been investigated. For change detection analysis, data manipulation in several steps involving preprocessing, processing and colour display have been carried out. Land-use changes have been detected by (a) image differencing, (b) image ratioing, and (c) differencing of NDVI images. It is inferred from the remote sensing images that extensive mining, establishment of communication networks, expansion of settlements, decrease in the vegetation cover etc., have remodelled the face of the JCF.     

Differences in monitoring vegetation dynamics between Moderate Resolution Imaging Spectroradiometer Collection 5 and Collection 6 vegetation index products on the Loess Plateau,China

Accurately monitoring vegetation dynamics on the Loess Plateau (LP) is critical for evaluating the benefits of ecological restoration projects. The Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation index (VI) product has been a primary data source for monitoring vegetation dynamics. However, MODIS Collection 5 (C5) VI products are known to be affected by sensor degradation, which has been addressed in the newly released MODIS Collection 6 (C6) VI products. Herein, we compared the spatiotemporal differences in vegetation dynamics between the Terra MODIS C5 and C6 data products and among different annual value retrieval methods for the LP during 2001–2016. Our results indicated a lower magnitude but a greener trend in the normalized difference vegetation index (NDVI), and areas with significant greening (p < 0.050) were found to increase by about 13%–16% from C5 to C6, depending on the retrieval method. Regions with either no particular trend or a downward trend in vegetation derived from the Terra-C5 NDVI mostly showed significant increasing trends based on the Terra-C6 NDVI. Moreover, the different retrieval methods also exhibited differences in the evaluation of vegetation dynamics, with the largest differences in terms of both magnitude and trend being identified with the annual maximum value method. This highlighted a compelling need to choose suitable methods in different regions for the retrieval of annual VI values, in order to facilitate more robust and comparable conclusions. Additionally, discrepancies also existed in the response of vegetation to climate variations between the Terra-C5 and C6 products for all three annual VI retrieval methods. Our findings, based on multiple products and analysis methods, may lead to improved understanding of both vegetation dynamics and their linkage to climate variables. The results suggest that caution be utilized when using only MODIS Terra-C5 products to evaluate vegetation dynamics and calibrate ecosystem models.     

Methods and tools supporting modular process design

Increasing competition in globalising markets requires effective means for development and production planning regarding innovative products. One efficient approach is an integrated product and process design, following the idea of concurrent engineering. Initialising process design at an early stage of product development allows for balancing the product characteristics and the process capabilities. In this connection, the architecture of process planning operations has to reflect the information certainty derived from the development process. At the Laboratory for Machine Tools and Production Engineering (WZL), methods and electronic data processing (EDP) tools for a continuous, product related process design in different planning and optimisation phases have been developed. The procedure and the efficiency of the developed methods and EDP-tools which are designed modularly are illustrated in the following paper.     

An assessment of validation techniques for estimating chlorophyll-a concentration from airborne multispectral imagery

Accurate estimates of chlorophyll-a levels in coastal waters are required for the assessment of waters potentially subject to eutrophication. Traditional laboratory analysis of water samples does not offer the required spatial or temporal density of sampling. Remote sensing methods have been suggested as a more appropriate representation of the variability in chlorophyll a concentration encountered in UK coastal waters. This paper examines the use of two established techniques for the calibration of airborne multispectral imagery to determine chlorophyll-a concentration: the blue/green ratio and Fluorescence Line Height methods. These methods have been developed for use in oceanic or Canadian coastal waters. The errors incurred in the use of these techniques for calibration of data of UK coastal waters have not previously been addressed. This paper describes a rigorous assessment of the errors incurred, allowing the limitations of the techniques to be established. This has resulted in recommendations for chlorophyll-a measurement in the coastal zone.     

Merging of IRS LISS III and PAN data—evaluation of various methods for a predominantly agricultural area

The present study evaluates six different methods of data merging using Indian Remote Sensing Satellite (IRS) panchromatic (PAN; high spatial resolution) and Linear Imaging Self-scanning Sensor (LISS) III (high spectral resolution) data for a predominantly agricultural area including a potato research farm in Jalandhar, Punjab, India. The methods used were intensity–hue–saturation (IHS), principal component substitution (PCS), high pass filter (HPF), Brovey, synthetic variable ratio (SVR) and cubic spline wavelet (CSW) technique. The LISS III data, which were of the same date as the PAN data, were registered to the PAN data by identifying ground control points (GCPs) and using the cubic convolution resampling method. Merged data products were generated using the above-mentioned six techniques. The merged products were evaluated on three aspects, i.e. visually, statistically and by comparing classification accuracy. The study could help to rank the suitability of various merging methods for agricultural land-use applications. The HPF, SVR and CSW merging methods were more accurate than the commonly used IHS, PCS and Brovey methods. The PCS was found to be least accurate among all.     

Information technology and structural change in the paper and pulp industry : Some technological, organizational and managerial implications

The paper and pulp industry has experienced a structural change during the last 15 years. The products have differentiated and completely new paper types have been introduced. The standard and classical products, like news print paper, are often specialized and customized versions from the standard product in order to fulfill the specific needs of different printing processes better. These changes have impacted on the production organization and structures of enterprises. The key companies have acquired more production capacity in different countries and have become really international companies. These trends are likely to continue. One of the technological driving forces has been the applying of information technology. The digital process automation systems as well as the mill wide information systems have been strategic tools and technological corner stones in the structural change. However, it is only technology which matters. Also new organisational forms as well as new implementation methods of systems are needed.     

Exploring artificial intelligence-based data fusion for conjoint analysis

Conjoint analysis is used to understand how consumers develop preferences for products or services, which encompass, as usual, multi-attributes and multi-attribute levels. Conjoint analysis has been one of the popular tools for multi-attribute decision-making problems on products and services for consumers over the last 30 years. It has also been used to market segmentation and optimal product positioning. In spite of its popularity and commercial success, a major weakness of conjoint analysis has been pointed such that respondents participating in conjoint experiment have to evaluate a number of hypothetical product profiles. To reduce the number of hypothetical products, this paper proposes a systematic method, called data fusion, and explores the usability of various data fusion techniques. The paper evaluates traditional data fusion (correlation-based), hierarchical Bayesian-based data fusion, and neural network-based data fusion.     

A quantitative correlation coefficient mining method for business intelligence in small and medium enterprises of trading business

Business intelligence based on data mining has been one of the popular and indispensable tools for identifying business opportunity in sales and marketing of new products. The traditional data mining methods based on association rules may be inadequate in completely uncovering the hidden patterns of sales based on transaction records. This paper presents a qualitative correlation coefficient mining method which is capable of uncovering hidden patterns of sales and market. Hence, a prototype business intelligence system (BIS) named correlation coefficient sales data mining system (CCSDMS) has been developed and successfully trial implemented in a selected reference site. A series of experiments have been conducted to evaluate the performance of the proposed system. The results generated by the BIS are compared with a well known market available data mining system. The proposed quantitative correlation coefficient mining method is found to possess higher accuracy, better computational effectiveness and higher predictive power. With the new approach, associations for product relations and customer periodic demands are revealed and this can help to leverage organizational marketing capital to enhance quality and speed of promotions as well as awareness of product relations.     

IKONOS全色与多光谱数据融合方法的比较研究

IKNOS-2给各个应用领域提供1 m的全色和4 m的多光谱数据,因此利用全色波段将4 m的多光谱数据融合为1 m的多光谱数据会充分利用二者的信息,提高目视和自动影像提取的类别精度。影像融合技术发展较快,成为遥感应用研究领域的重要主题。在遥感领域应用较多的融合方法有IHS变换、主成分分析、Brovey(颜色归一化)变换、小波变换以及最近发展修改的合成变量比值变换。从光谱质量和空间信息角度对融合方法进行了比较研究,筛选出适合IKNOS融合方法,从各种特征信息提取和自动分类角度出发,合成变量比值变换融合方法光谱退化最小,同时也较高程度地保持了高几何分辨率全色的空间信息。     

基于FTRL和XGBoost算法的产品故障预测模型

随着智能化设备的日益更新和计算机储存数据能力的提升,制造业企业在其产品制造过程中产生了大量的流水线数据,如何充分利用这些数据一直是工业界的一个难题.本文根据制造业企业的真实大规模生产数据,通过对其进行细致的探索性数据分析,建立了一种基于FTRL和XGBoost算法的二分类产品故障预测模型,并根据适用于非平衡数据集的MCC （Matthews Correlation Coefficient）评价指标采用交叉验证方法对其进行优化.实验结果表明,该模型对于大规模（不仅样本量大,特征量也很大）正负样本非平衡的生产流水线数据集具有运行效率高,故障预测精度高的效果.基于此模型我们可以构建更智能的产品故障检测系统,有效降低企业运营成本的同时也带来了可观的利润增长.     

Comparison between linear genetic programming and M5 tree models to predict flow discharge in compound channels

There are many studies on the hydraulic analysis of steady uniform flows in compound open channels. Based on these studies, various methods have been developed with different assumptions. In general, these methods either have long computations or need numerical solution of differential equations. Furthermore, their accuracy for all compound channels with different geometric and hydraulic conditions may not be guaranteed. In this paper, to overcome theses limitations, two new and efficient algorithms known as linear genetic programming (LGP) and M5 tree decision model have been used. In these algorithms, only three parameters (e.g., depth ratio, coherence, and ratio of computed total flow discharge to bankfull discharge) have been used to simplify its applications by hydraulic engineers. By compiling 394 stage-discharge data from laboratories and fields of 30 compound channels, the derived equations have been applied to estimate the flow conveyance capacity. Comparison of measured and computed flow discharges from LGP and M5 revealed that although both proposed algorithms have considerable accuracy, LGP model with R ² = 0.98 and RMSE = 0.32 has very good performance.     

Guest Editors' Introduction: Progress in Test Compression

This special issue represents a snapshot of the progress in test compression, a key strategy for dealing with rapidly growing test data volume. Test compression involves encoding test data in a compressed form so that less data needs to be transferred, thereby reducing test time and the need for tester memory. A wide variety of test compression techniques have been developed, both for compressing test vectors and compressing output responses. In recent years, several researchers and companies have developed compression methods and products that achieve significant amounts of compression. These new methods have extended the life of legacy ATE and have been synergistic with the need for additional tests to detect the defects arising in nanometer designs. Thus, test compression continues to be a very active area.     

Learning low-rank kernel matrices for constrained clustering

Constrained clustering methods (that usually use must-link and/or cannot-link constraints) have been received much attention in the last decade. Recently, kernel adaptation or kernel learning has been considered as a powerful approach for constrained clustering. However, these methods usually either allow only special forms of kernels or learn non-parametric kernel matrices and scale very poorly. Therefore, they either learn a metric that has low flexibility or are applicable only on small data sets due to their high computational complexity. In this paper, we propose a more efficient non-linear metric learning method that learns a low-rank kernel matrix from must-link and cannot-link constraints and the topological structure of data. We formulate the proposed method as a trace ratio optimization problem and learn appropriate distance metrics through finding optimal low-rank kernel matrices. We solve the proposed optimization problem much more efficiently than SDP solvers. Additionally, we show that the spectral clustering methods can be considered as a special form of low-rank kernel learning methods. Extensive experiments have demonstrated the superiority of the proposed method compared to recently introduced kernel learning methods.     

实用软件质量管理方法

随着信息技术在军民用产品中的广泛应用，软件已成为大多数产品的重要组成部分，如何提高软件质量，使软件更好地服务于各种应用需要，已成为各行各业广泛关注的课题。本文针对软件的特点和软件研制的实践情况，提出了软件质量管理实用方法，该方法通过实际应用并取得较好的效果。