排序方式: 共有53条查询结果,搜索用时 15 毫秒
1.
2.
In this paper, the problem of residual variance estimation is examined. The problem is analyzed in a general setting which covers non-additive heteroscedastic noise under non-iid sampling. To address the estimation problem, we suggest a method based on nearest neighbor graphs and we discuss its convergence properties under the assumption of a Hölder continuous regression function. The universality of the estimator makes it an ideal tool in problems with only little prior knowledge available. 相似文献
3.
El‐Hajjaji Soundous Grard Amaury De Laubier Juliette Di Tanna Sybille Lain Aurlie Patz Viviane Sindic Marianne 《International Journal of Dairy Technology》2019,72(3):466-471
This study aims to describe the procedures and practices used in local production of raw milk butter. The demand for local products is increasing; hence, there is a need to describe the practices used in the artisanal production of raw milk butter. Therefore, a survey of 147 raw milk butter producers was carried out. The results from the survey indicate that there is not one single way to produce butter at artisanal level. In terms of maturation, six temperature sequences were distinguished. Attention is required at every step of production starting from breeding. 相似文献
4.
Similarity functions are a fundamental component of many learning algorithms. When dealing with string or tree-structured data, measures based on the edit distance are widely used, and there exist a few methods for learning them from data. However, these methods offer no theoretical guarantee as to the generalization ability and discriminative power of the learned similarities. In this paper, we propose an approach to edit similarity learning based on loss minimization, called GESL. It is driven by the notion of (?,??,??)-goodness, a theory that bridges the gap between the properties of a similarity function and its performance in classification. Using the notion of uniform stability, we derive generalization guarantees that hold for a large class of loss functions. We also provide experimental results on two real-world datasets which show that edit similarities learned with GESL induce more accurate and sparser classifiers than other (standard or learned) edit similarities. 相似文献
5.
Buzzi Maria Claudia Buzzi Marina Leporini Barbara Trujillo Amaury 《Multimedia Tools and Applications》2017,76(4):5141-5169
Multimedia Tools and Applications - We present an analysis of how visually impaired people perform gestures on touch-screen smartphones and report their preferences, explaining the procedure and... 相似文献
6.
Amaury Delamarre Myriam Paire Jean‐Franois Guillemoles Laurent Lombez 《Progress in Photovoltaics: Research and Applications》2015,23(10):1305-1312
We investigate photoluminescence and electroluminescence (PL and EL) emission images from Cu(In,Ga)Se2‐based solar cells by means of a Hyperspectral Imager. Using the generalized Planck's law, maps of the effective quasi‐Fermi level splitting Δμeff in absolute values are obtained. A good agreement is found between the spatially averaged splitting in PL and the global open‐circuit voltage. However, from a local carrier transport discussion, we conclude that the equality does not hold locally. The spatial variations are rather attributed to local depth variations of the quasi‐Fermi level splitting due to material properties spatial fluctuations. By comparing PL and EL emissions, we discuss qualitatively the local effective lifetimes and collection efficiencies. Copyright © 2014 John Wiley & Sons, Ltd. 相似文献
7.
Rice, a cereal widely used in Venezuela for human and animal nutrition, is susceptible to aflatoxin contamination in the field and during storage. Therefore, the goal of this research was the evaluation of the efficacy and permanence of the ammoniation process through high pressure/high temperature (HP/HT) and atmospheric pressure/moderate temperature (AP/MT) conditions applied to rice samples artificially contaminated with aflatoxin B1. For this purpose a 2k design was drawn up considering the temperature, the rice moisture and the process time as variables. Under both sets of conditions, aflatoxin B1 concentration was reduced in a range of 90% to 100%. After in vitro stomach digestion simulation, toxin reversion ranged from 0% to 19%. In conclusion, the process efficacy and permanence were achieved through the use of high temperature and long process time for both sets of conditions (HP/HT and AP/MT), respectively. 相似文献
8.
3D object recognition based on a geometrical topology model and extreme learning machine 总被引:1,自引:1,他引:0
In this paper, one geometrical topology hypothesis is present based on the optimal cognition principle, and the single-hidden layer feedforward neural network with extreme learning machine (ELM) is used for 3D object recognition. It is shown that the proposed approach can identify the inherent distribution and the dependence structure for each 3D object along multiple view angles by evaluating the local topological segments with a dipole topology model and developing the relevant mathematical criterion with ELM algorithm. The ELM ensemble is then used to combine the individual single-hidden layer feedforward neural network of each 3D object for performance improvements. The simulation results have shown the excellent performance and the effectiveness of the developed scheme. 相似文献
9.
10.
Summary: Polystyrene (PS)/epoxy‐amine (DGEBA‐MDEA) is a thermoplastic/thermoset precursor blend which is miscible at high temperature (177 °C), and which phase separates under the polymerization of the epoxy‐amine system. Previous studies have shown that the morphology of this blend polymerized under shear is coarse and irregular because the dispersed epoxy‐amine domains coalesce before they gel. Several styrene‐methyl methacrylate and a styrene‐butadiene‐styrene block copolymers have been added to the PS/DGEBA‐MDEA 60/40 blend in order to limit the coalescence and thus obtain a finer morphology. Two of the copolymers used were reactive either with the epoxy or with the amine. It was shown that the addition of 15 wt.‐% of non reactive copolymer had a positive but limited effect on the size of the final epoxy‐amine particles. The copolymer remained at the interface at the early stages of the polymerization. However, it was pulled out by the shear forces around the gel point of the epoxy domains. Most of the non reactive copolymer was present in the shape of micelles at the end of the process. On the other hand, the reactive copolymers were able to establish covalent bonds with the epoxy‐amine drops and hence were not extracted at all. Consequently they allowed the decrease the size of the particles by a factor of 15. Despite this, the observation of the morphology at different stages of the polymerization has revealed that the copolymer moved at the interface of the epoxy domains during the collision of two droplets. The movements of fluids into the epoxy domains pushed the copolymer out of the inter‐droplet zone so that it could not prevent the drainage of the liquid film between the droplets and consequently their coalescence.