基于模糊支持向量机的小批量生产质量智能预测方法

在分析比较目前常用的智能工序预测技术及其特点的基础上,提出一种适合小批量生产过程的质量智能预测模型,并给出了相应的预测过程和算法.由于该模型中以模糊支持向量机(FLS-SVM)技术为智能核心,一方面较好的解决小样本学习问题,避免了人工神经网络等智能方法在对小批量生产过程质量进行预测时所表示出的过学习、泛化能力弱等缺点.另一方面,通过隶属度函数对样本进行模糊化,达到样本优化选择,实现历史数据“重近轻远”的预测效果.通过对具体加工过程的预测实验,并与其它几种常见预测方法效果进行对比,说明本文方法实现容易,建模速度快,小样本的泛化能力强,为实现小批量加工过程的在线质量预测与控制提供可行的思路.     

应用支持向量机预测公交车运行时间

采用一种新颖的神经网络-支持向量机(SVM),来预测公交车的到站时间,其目的是要验证SVM在运行时间预测领域的可行性.该模型采用了时间段、天气、路段以及当前路段的运行时间和下一路段的最新运行时间5个输入变量.最后,应用大连市开发区4路公交线对该模型进行了校验,并得到若干结论.     

质子交换膜燃料电池混合建模研究

支撑向量机（SVM）理论完备,泛化能力强,很适合对燃料电池建模。但是建立高维SVM模型需要大量的实验数据,为了克服这一困难,使用SVM和压力增量机理模型相结合的方法对质子交换膜燃料电池进行混合建模：SVM模型只考虑电流和温度对电压的影响,而压力增量模型则在此基础上预测阴极压力和阳极压力对电压的影响。混合后的模型能够预测不同电流、温度、阴极压力和阳极压力下的输出电压。结果表明这种方法建立的数学模型误差小于1.6%,能够达到很好的拟合精度。     

最小二乘支持向量机的预优共轭梯度法

针对Suyken等人提出的最小二乘支持向量机的共轭梯度法在输入样本的个数较大时,需要求解高阶线性方程组这一缺陷,提出了一种新算法。该算法利用分块矩阵的思想将该高阶线性方程组系数矩阵降阶,为了提高收敛速度,克服数值的不稳定性,采用条件预优共轭梯度法求解低阶的线性方程组。通过仿真试验证明用本文方法训练最小二乘支持向量机比共轭梯度法的训练速度提高了将近一倍。     

Estimation of underdetermined mixing matrix based on support vector machine

In underdetermined blind source separation (BSS), a novel algorithm based on extended support vector machine (SVM) is proposed to estimate the mixing matrix in this paper, including the number of the active sources. Instead of traditional clustering algorithms, it mainly takes the modulus of observations and the number in each direction of arrival, without any prior knowledge about the sources except for sparsity, and it is not sensitive to the initial values. Simulations are given to illustrate availability and robustness of our algorithm.     

What can we learn from the fifties?

Economists have increasingly elicited probabilistic expectations from survey respondents. Subjective probabilistic expectations show great promise to improve the estimation of structural models of decision making under uncertainty. However, a robust finding in these surveys is an inappropriate heap of responses at “50%,” suggesting that some of these responses are uninformative. The way these 50s are treated in the subsequent analysis is of major importance. Taking the 50s at face value will bias any aggregate statistics. Conversely, deleting them is not appropriate if some of these answers do convey some information. Furthermore, the attention of researchers is so focused on this heap of 50s that they do not consider the possibility that other answers may be uninformative as well. This paper proposes to take a fresh look at these questions using a new method based on weak assumptions to identify the informativeness of an answer. Applying the method to probabilistic expectations of equity returns in three waves of the Survey of Economic Expectations in 1999–2001, I find that: (i) at least 65% of the 50s convey no information at all; (ii) it is the answer most often provided among the answers identified as uninformative; (iii) but even if the 50s are a major contributor to noise, they represent at best 70% of the identified uninformative answers. These findings have various implications for survey design.     

Microhabitat diversity of Svalbard Bryozoa

Bryozoans are one of the major macrofaunal groups of the high polar regions. Here we present data on the nature of bryozoan assemblages in the Svalbard Archipelago sampled over 6 years between 1997 and 2002. Samples were collected with use of Van Veen grab, dredge and Scuba diving at depths ranging from 0 to 268?m. Among examined material (5026 items) bryozoans were present on one type of abiotic (stones) and 40 types of biotic substrata. The biotic substrata we investigated included algae (12 taxa) and invertebrates (28 taxa: Hydrozoa, Gastropoda, Bivalvia, Polychaeta, Crustacea, Ascidiacea, other Bryozoa). Both a priori ANOSIM (Global R?=?0.512, P?=?0.001) and cluster analyses reveal differences between the assemblages on different types of substrata. Cluster analyses distinguish two distinct groups of substrata. The first (stable) included stones, molluscs and Balanus balanus whilst the second (flexible) consisted of algae, hydrozoans and bryozoans. Bryozoan assemblages which colonized different substrata, varied in terms of species composition and their dominance. Bryozoan species were categorized as generalists, low‐specificity epibiotic species or locally abundant background species. No species were found to be substratum‐specific to any type. Among substrata with the richest bryozoan fauna were stones (156 taxa), Chlamys islandica (68) and Balanus balanus (62). There were clear patterns in the distribution of substrata colonized by bryozoans within the fjord system: those shallow near the mouth were rich whilst those in deeper water were depauperate. We consider environmental gradients such as siltation and processes related to depth to be responsible for such patterns.     

On the natural history of the South American pepper frog,Leptodactylus labyrinthicus (Spix, 1824) (Anura: Leptodactylidae)

Some species of Leptodactylus of the L. pentadactylus group lay their eggs outside water but the tadpoles need to reach water to complete the larval phase; other species complete development in terrestrial nests. Here we present details of the reproduction of L. labyrinthicus in south‐eastern Brazil. The proportion of tadpoles and trophic eggs in aged egg clutches was determined, as well as the growth of the tadpoles while in the nest. The gut contents of tadpoles that were in egg clutches of frogs were analysed. Adult males did not differ from females in size and had hypertrophied forearms and an enlarged spine on the thumb. Reproduction was initiated with the first rains of August/September and extended to mid‐January. Calling and spawning occurred at permanent or temporary water bodies. The foam nests were built in excavated basins outside of, but close to the water. The male determined the place of the basin construction; after amplexus, the female completed the excavation. The amplexus was axillary. One female spent the day after spawning in the foam. The eggs were pale grey, the yolk averaging 2.3?mm in diameter. The mean number of eggs was 2101 per egg clutch. The number of tadpoles in individual nests varied between 0.05% and 11.40% in relation to the total laid eggs. The tadpoles entered water when rains flooded the basin. The tadpoles grew to 12 times the weight of an individual egg while in the nest; no nesting tadpole was beyond stage 25. The longest time we followed tadpoles in a nest was 25 days. Tadpoles were found preying upon eggs of three other frog species and upon conspecific eggs. Males fought by grasping each other in a belly‐to‐belly position; the powerful arms and the thumb spines represent weapons. Even though males can reach maturity in the season following birth, small size would prevent them from establishing their own territory. All the species of the L. pentadactylus group may build their foam nests within excavated basins. The basins may protect the eggs and embryos from cannibalistic tadpoles and may have an anti‐desiccation effect. In order to produce trophic eggs, the female may delay laying additional unfertilized eggs until after the male has abandoned the foam nest. Anuran eggs represent an important food item for tadpoles after they leave the nest.     

Geographical variation in the golden‐striped salamander,Chioglossa lusitanica Bocage, 1864 and the description of a newly recognized subspecies

Morphological and genetic data for the Iberian golden‐striped salamander, Chioglossa lusitanica, demonstrate the existence of two groups with southern and northern ranges, connected by a zone of intergradation in central Portugal. Because reproductive isolation between them is incomplete we consider the groups to be subspecies. The type locality of C. lusitanica (Buçaco near Lousã) is situated inside the mixed zone. This necessitates identification of the nominotypical subspecies. We sequenced a fragment of mitochondrial DNA from one of the species' syntypes and we determined what position over a latitudinal transect maximizes the morphological discrimination between the groups. Both approaches indicate that C. lusitanica from Buçaco represents the southern subspecies. A new subspecies of C. lusitanica is described from a northern locality (Valongo near Porto in north‐western Portugal). A lectotype is designated for Chioglossa lusitanica.     

Forecasting the 10‐year US Treasury rate

This study compares the performance of two forecasting models of the 10‐year Treasury rate: a random walk (RW) model and an augmented‐autoregressive (A‐A) model which utilizes the information in the expected inflation rate. For 1993–2008, the RW and A‐A forecasts (with different lead times and forecast horizons) are generally unbiased and accurately predict directional change under symmetric loss. However, the A‐A forecasts outperform the RW, suggesting that the expected inflation rate (as a leading indicator) helps improve forecast accuracy. This finding is important since bond market efficiency implies that the RW forecasts are optimal and cannot be improved. Copyright © 2009 John Wiley & Sons, Ltd.