• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.6
• /
• pp.2527-2545
• /
• 2016

Measuring the similarity of given samples is a key problem of recognition, clustering, retrieval and related applications. A number of works, e.g. kernel method and metric learning, have been contributed to this problem. The challenge of similarity learning is to find a similarity robust to intra-class variance and simultaneously selective to inter-class characteristic. We observed that, the similarity measure can be improved if the data distribution and hidden semantic information are exploited in a more sophisticated way. In this paper, we propose a similarity learning approach for retrieval and recognition. The approach, termed as LDA-FEK, derives free energy kernel (FEK) from Latent Dirichlet Allocation (LDA). First, it trains LDA and constructs kernel using the parameters and variables of the trained model. Then, the unknown kernel parameters are learned by a discriminative learning approach. The main contributions of the proposed method are twofold: (1) the method is computationally efficient and scalable since the parameters in kernel are determined in a staged way; (2) the method exploits data distribution and semantic level hidden information by means of LDA. To evaluate the performance of LDA-FEK, we apply it for image retrieval over two data sets and for text categorization on four popular data sets. The results show the competitive performance of our method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.8
• /
• pp.2895-2912
• /
• 2014

Content-based image retrieval has been the most important technique for managing huge amount of images. The fundamental yet highly challenging problem in this field is how to measure the content-level similarity based on the low-level image features. The primary difficulties lie in the great variance within images, e.g. background, illumination, viewpoint and pose. Intuitively, an ideal similarity measure should be able to adapt the data distribution, discover and highlight the content-level information, and be robust to those variances. Motivated by these observations, we in this paper propose a probabilistic similarity learning approach. We first model the distribution of low-level image features and derive the free energy kernel (FEK), i.e., similarity measure, based on the distribution. Then, we propose a learning approach for the derived kernel, under the criterion that the kernel outputs high similarity for those images sharing the same class labels and output low similarity for those without the same label. The advantages of the proposed approach, in comparison with previous approaches, are threefold. (1) With the ability inherited from probabilistic models, the similarity measure can well adapt to data distribution. (2) Benefitting from the content-level hidden variables within the probabilistic models, the similarity measure is able to capture content-level cues. (3) It fully exploits class label in the supervised learning procedure. The proposed approach is extensively evaluated on two well-known databases. It achieves highly competitive performance on most experiments, which validates its advantages.

• Animal Bioscience
• /
• v.36 no.5
• /
• pp.768-775
• /
• 2023

Objective: Oilseeds and nut co-products can be used as alternative feed ingredients in animal diets because they may have a lower cost than traditional ingredients. A study was, therefore, conducted to determine the chemical composition of copra, palm kernel, and nut co-products from South-East Asia or Australia. The hypothesis that country of production influences nutritional composition was tested. Methods: Oilseed meals included 2 copra expellers, 3 copra meals, and 12 palm kernel expellers. One source of almond hulls and cashew nut meal were also used. Samples were obtained from suppliers located in South-East Asia or Australia. All samples were analyzed for dry matter, gross energy, nitrogen, amino acids (AA), acid-hydrolyzed ether extract (AEE), ash, minerals, insoluble dietary fiber, and soluble dietary fiber. Copra and nut co-products were also analyzed for total starch and sugars. Results: Copra expellers had greater (p<0.05) concentrations of dry matter and AEE compared with copra meal. However, copra meal had greater (p<0.05) concentrations of total dietary fiber (soluble and insoluble) and copper than copra expellers. Palm kernel expellers from Indonesia had greater (p<0.05) concentration of histidine and tyrosine compared with palm kernel expellers from Vietnam. Almond hulls was high in dietary fiber, but also contained free glucose and fructose, whereas cashew nut meal was high in AEE, but low in all free sugars. Conclusion: Copra expellers have greater concentration of AEE, but less concentration of total dietary fiber when compared with copra meal, and except for a few AA, no differences in nutrient composition of palm kernel expellers produced in Indonesia or Vietnam were detected. According to the chemical composition of nut co-products, cashew nut meal may be more suitable for non-ruminant diets than almond hulls.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.5 no.4
• /
• pp.513-528
• /
• 2013

A floating Oscillating Water Column (OWC) wave energy converter, a Backward Bent Duct Buoy (BBDB), was simulated using a state-of-the-art, two-dimensional, fully-nonlinear Numerical Wave Tank (NWT) technique. The hydrodynamic performance of the floating OWC device was evaluated in the time domain. The acceleration potential method, with a full-updated kernel matrix calculation associated with a mode decomposition scheme, was implemented to obtain accurate estimates of the hydrodynamic force and displacement of a freely floating BBDB. The developed NWT was based on the potential theory and the boundary element method with constant panels on the boundaries. The mixed Eulerian-Lagrangian (MEL) approach was employed to capture the nonlinear free surfaces inside the chamber that interacted with a pneumatic pressure, induced by the time-varying airflow velocity at the air duct. A special viscous damping was applied to the chamber free surface to represent the viscous energy loss due to the BBDB's shape and motions. The viscous damping coefficient was properly selected using a comparison of the experimental data. The calculated surface elevation, inside and outside the chamber, with a tuned viscous damping correlated reasonably well with the experimental data for various incident wave conditions. The conservation of the total wave energy in the computational domain was confirmed over the entire range of wave frequencies.

• KEPCO Journal on Electric Power and Energy
• /
• v.4 no.2
• /
• pp.101-105
• /
• 2018

In this paper, the kernel regression model is applied for the case study of gas turbine abnormal state analysis. In addition to vibration analysis at the remote site, the kernel regression model technique can is useful for analyzing abnormal state of rotor vibration signals of gas turbine in power plant. In monitoring based on data-driven techniques correlated measurements, the fault free training data of shaft vibration obtained during normal operations of gas turbine are used to develop a empirical model based on auto-associative kernel regression. This data-driven model can be used to predict virtual measurements, which are compared with real-time data, generating residuals. Any faults in the system may cause statistically abnormal changes in these residuals and could be detected. As the result, the kernel regression model provides information that can distinguish anomalies such as sensor failure in a shaft vibration signal.

• KIEE International Transaction on Electrical Machinery and Energy Conversion Systems
• /
• v.4B no.4
• /
• pp.180-183
• /
• 2004

This paper presents a description of the point collocation method and its application to the electromagnetic field computation. The interpolation scheme is based on the fast moving least square reproducing kernel approximation. In the method, the integration cell is not required and the essential boundary conditions can be enforced directly. Numerical simulations on 1-D and 2-D problems are carried out to validate the method. It is found that computational efficiency is higher than the general mesh-free methods.

• Asian-Australasian Journal of Animal Sciences
• /
• v.15 no.9
• /
• pp.1273-1277
• /
• 2002

The effect of urea treatment on chemical composition of neem seed kernel cake (NSKC) was investigated by soaking the cake in 2.1% urea solution (1.2 l $kg^{-1}$ NSKC) for five days. The effect on utilisation of nutrients by replacing groundnut cake (GNC) (30%) with urea-treated neem seed kernel cake (UTNSKC) (33%) in a concentrate mixture fed to meet 70% of the protein requirements of lambs (8 males and 8 females), was monitored in a digestibility study. Following urea treatment of NSKC only 9.5% of urea was hydrolysed and the crude protein content of the cake was increased by 6.65%. The tannin content in depulped neem seeds was 37% catechin equivalent. Whereas feeding UTNSKC had no effect on intake of dry matter (72.5 vs 66.3 g/kg $BW^{0.75}day^{-1}$) and digestibility of crude fibre (41.3 vs 43.4%), the cake depressed (p<0.01) the percent digestibility of dry matter (63.7 vs 70.2), crude protein (63.2 vs 70.2), nitrogen free extract (73.8 vs 80.5) and gross energy (64.3 vs 69.1). Digestibility of ether extract (75.8 vs 70.9%) was higher (p<0.05) in animals offered UTNSKC. The nutritive value of the composite ration consumed by lambs offered UTNSKC was lower (p<0.01) in terms of total digestible nutrients (64.7 vs 70.2%) and digestible energy (2.8 vs 3.0 Kcal/g DM). Intake of digestible energy (199.8 vs 194.1 Kcal/kg $BW^{0.75}day^{-1}$) and retention of nitrogen (7.53 vs 8.23 g $day^{-1}$) and calcium (2.12 vs 1.84 g $day^{-1}$) were comparable on the 2 rations. Animals fed UTNSKC retained less (p<0.01) phosphorus (0.37 vs 1.05 g $day^{-1}$). The results indicate that urea treatment increased the protein level of NSKC whereas feeding the treated cake as a replacement of GNC, lowered the digestibility of nutrients and retention of phosphorus.

• Asian-Australasian Journal of Animal Sciences
• /
• v.26 no.4
• /
• pp.537-544
• /
• 2013

This study examined whether pre-treating palm kernel expeller (PKE) with exogenous enzyme would degrade its fiber content; thus improving its metabolizable energy (ME), growth performance, villus height and digesta viscosity in broiler chickens fed diets containing PKE. Our results showed that enzyme treatment decreased (p<0.05) hemicellulose and cellulose contents of PKE by 26.26 and 32.62%, respectively; and improved true ME (TME) and its nitrogen corrected value ($TME_n$) by 38% and 33%, respectively, compared to the raw sample. Average daily gain (ADG), feed intake and feed conversion ratio (FCR) of chickens fed on different dietary treatments in the grower period were not significantly different. Although there was no difference in feed intake (p>0.05) among treatment groups in the finisher period, ADG of chickens in the control (PKE-free diet) was higher (p<0.05) than in all treatment groups fed either 20 or 30% PKE, irrespective of with or without enzyme treatment. However, ADG of birds fed with 20% PKE was higher than those fed with 30% PKE. The FCR of chickens in the control was the lowest (2.20) but not significantly different from those fed 20% PKE diets while birds in the 30% PKE diets recorded higher (p>0.05) FCR. The intestinal villus height and crypt depth (duodenum, jejunum and ileum) were not different (p>0.05) among treatments except for duodenal crypt depth. The villus height and crypt depth of birds in enzyme treated PKE diets were higher (p<0.05) than those in the raw PKE groups. Viscosity of the intestinal digesta was not different (p>0.05) among treatments. Results of this study suggest that exogenous enzyme is effective in hydrolyzing the fiber (hemicellulose and cellulose) component and improved the ME values of PKE, however, the above positive effects were not reflected in the growth performance in broiler chickens fed the enzyme treated PKE compared to those received raw PKE. The results suggest that PKE can be included up to 5% in the grower diet and 20% in the finisher diet without any significant negative effect on FCR in broiler chickens.

• Proceedings of the KSME Conference
• /
• 2001.06a
• /
• pp.680-685
• /
• 2001

In this paper, RKPM is extended for solving moderately thick and thin structures. General Timoshenko beam and Mindlin plate theory are used far formulation. Shear locking is the main difficulty in analysis of these kinds of structures. Shear relaxation factor, which is formulated using the difference between bending and shear strain energy, is introduced to overcome shear locking. Analysis results obtained reveal that RKPM using introduced method is free of locking and very effectively applicable to deeply as well as shallowly beams and plates.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.12 no.5
• /
• pp.73-79
• /
• 2003

In this paper, RKPM is extended for solving moderately thick and thin beams. General Timoshenko beam theory is used for formulation. Shear locking is the main difficulty in analysis of these kinds of structures. Shear relaxation factor, which is formulated using the difference between bending and shear strain energy, and corrected shear rigidity are introduced to overcome shear locking. Analysis results obtained reveal that RKPM using introduced methods is free of locking and very effectively applicable to deep beams as well as shallow beams.