• Communications for Statistical Applications and Methods
• /
• v.11 no.3
• /
• pp.475-484
• /
• 2004

This paper consider the testing problem of grouped heteroscedasticity in the linear regression model. We provide the Lagrange Multiplier(LM), Wald, Likelihood Ratio (LR) test statistis for testing of grouped heteroscedasticity. Monte Carlo experiments are conducted to study the performance of these tests.

• KIISE Transactions on Computing Practices
• /
• v.22 no.7
• /
• pp.315-319
• /
• 2016

Recurrent neural network based language models (RNN LM) have shown improved results in language model researches. The RNN LMs are limited to post processing sessions, such as the N-best rescoring step of the wFST based speech recognition. However, it has considerable vocabulary problems that require large computing powers for the LM training. In this paper, we try to find the 1st pass N-gram model using word embedding, which is the simplified deep neural network. The class based language model (LM) can be a way to approach to this issue. We have built class based vocabulary through word embedding, by combining the class LM with word N-gram LM to evaluate the performance of LMs. In addition, we propose that part-of-speech (POS) tagging based LM shows an improvement of perplexity in all types of the LM tests.

• ETRI Journal
• /
• v.38 no.3
• /
• pp.487-493
• /
• 2016

Two new methods are proposed for an unsupervised adaptation of a language model (LM) with a single sentence for automatic transcription tasks. At the training phase, training documents are clustered by a method known as Latent Dirichlet allocation (LDA), and then a domain-specific LM is trained for each cluster. At the test phase, an adapted LM is presented as a linear mixture of the now trained domain-specific LMs. Unlike previous adaptation methods, the proposed methods fully utilize a trained LDA model for the estimation of weight values, which are then to be assigned to the now trained domain-specific LMs; therefore, the clustering and weight-estimation algorithms of the trained LDA model are reliable. For the continuous speech recognition benchmark tests, the proposed methods outperform other unsupervised LM adaptation methods based on latent semantic analysis, non-negative matrix factorization, and LDA with n-gram counting.

• Journal of the Korean Statistical Society
• /
• v.36 no.4
• /
• pp.447-456
• /
• 2007

On the basis of marginal likelihood of the residual vector which is free of nuisance mean parameters, we propose new Lagrange Multiplier seasonal unit root tests in seasonal autoregressive process. The limiting null distribution of the tests is the standardized ${\chi}^2-distribution$. A Monte-Carlo simulation shows the new tests are more powerful than the tests based on the ordinary least squares (OLS) estimator, especially for large number of seasons and short time spans.

• International conference on construction engineering and project management
• /
• 2017.10a
• /
• pp.62-72
• /
• 2017

In this study, liner unit root tests and panel unit root tests to the ratio of city to regional house price were applied to examine the ripple effects across 28 cities in the Yangtze River Delta region. Then invert LM unit root tests with two structural breaks for 10 representative cities were conducted. The results showed that there is overwhelming evidence of the existence of ripple effect in the Yangtze River Delta region, while segmentation is restricted to a small group of cities in which there is no long-run relationship with the Yangtze River Delta region average; compared to no- and one-break case, there is overwhelming evidence of a ripple effect with the LM test with two structural breaks. Furthermore, the results of the Granger causality test showed that changes in house prices in Shanghai, Nanjing and Hangzhou have led to changes in house prices in other cities. The findings of this research make certain contributions to the improvements of research system of ripple effect among regional house prices in the Yangtze River Delta Region,and could be referenced by other markets of other cities.

• Journal of Korean Medicine for Obesity Research
• /
• v.13 no.1
• /
• pp.17-23
• /
• 2013

Objectives: This study was performed to evaluate the effects of fermented Samjung-hwan (SJH) extracts on weight, serum lipids and blood glucose. Methods: SJH was fermented using three different probiotic bacterial strains (Lactobacillus plantarum [LP], Leuconostoc mesenteroides [LM], Bifidobacterium longum [BL]) separately. Thirty-six rats were divided into normal, control (high fat diet), SJH-UF (high fat diet+unfermented SJH 200 mg/kg), SJH-LP (high fat diet+LP fermented SJH 200 mg/kg), SJH-LM (high fat diet+LM fermented SJH 200 mg/kg) and SJH-BL (high fat diet+BL fermented SJH 200 mg/kg). For 8 weeks later, we examined body weight, total cholesterol, high-density lipoprotein (HDL)-cholesterol and blood glucose. Results: The control group showed significantly increased weight gain compared with normal group and SJH-LP and BL groups had less weight gain than control group, significantly. In the lipid serum tests, control group showed significantly increased total cholesterol levels compared with normal group and only SJH-LP represented decreased total cholesterol levels compared with control group. However there was no significant change in the HDL-cholesteol levels. In the blood glucose tests, that of control group significantly incereased more than that of normal group, SJH-BL showed significantly decreased blood glucose levels compared with control group. Conclusions: SJH-LP, SJH-BL showed weight control effect, SJH-LP decreased TC and SJH-BL reduced blood glucose.

• Geomechanics and Engineering
• /
• v.28 no.4
• /
• pp.385-396
• /
• 2022

The accurate determination of the subgrade reaction modulus (K_s) of soil is an important factor for geotechnical engineers. This study estimated the K_s of soft soil improved with floating deep cement mixing (DCM) columns. A novel prediction model was developed that emphasizes the accuracy of identifying the most significant parameters of K_s. Several multi-layer perceptron (MLP) models that were trained using the Levenberg Marquardt (LM) backpropagation method were developed to estimate K_s. The models were trained using a reliable database containing the results of 36 physical modelling tests. The input parameters were the undrained shear strength of the DCM columns, undrained shear strength of soft soil, area improvement ratio and length-to-diameter ratio of the DCM columns. Grey wolf optimization (GWO) was coupled with the MLPs to improve the performance indices of the MLPs. Sensitivity tests were carried out to determine the importance of the input parameters for prediction of K_s. The results showed that both the MLP-LM and MLP-GWO methods showed high ability to predict K_s. However, it was shown that MLP-GWO (R = 0.9917, MSE = 0.28 (MN/m²/m)) performed better than MLP-LM (R =0.9126, MSE =6.1916 (MN/m²/m)). This proves the greater reliability of the proposed hybrid model of MLP-GWO in approximating the subgrade reaction modulus of soft soil improved with floating DCM columns. The results revealed that the undrained shear strength of the soil was the most effective factor for estimation of K_s.

• Steel and Composite Structures
• /
• v.39 no.3
• /
• pp.319-335
• /
• 2021

This study proposes a new and highly-accurate artificial intelligence model, namely ANN-IP, which combines an interior-point (IP) algorithm and artificial neural network (ANN), to improve the axial compression capacity prediction of elliptical concrete-filled steel tubular (CFST) columns. For this purpose, 145 tests of elliptical CFST columns extracted from the literature are used to develop the ANN-IP model. In this regard, axial compression capacity is considered as a function of the column length, the major axis diameter, the minor axis diameter, the thickness of the steel tube, the yield strength of the steel tube, and the compressive strength of concrete. The performance of the ANN-IP model is compared with the ANN-LM model, which uses the robust Levenberg-Marquardt (LM) algorithm to train the ANN model. The comparative results show that the ANN-IP model obtains more magnificent precision (R² = 0.983, RMSE = 59.963 kN, a20 - index = 0.979) than the ANN-LM model (R² = 0.938, RMSE = 116.634 kN, a20 - index = 0.890). Finally, a new Graphical User Interface (GUI) tool is developed to use the ANN-IP model for the practical design. In conclusion, this study reveals that the proposed ANN-IP model can properly predict the axial compression capacity of elliptical CFST columns and eliminate the need for conducting costly experiments to some extent.

• Ocean Systems Engineering
• /
• v.8 no.4
• /
• pp.427-439
• /
• 2018

Floating Production Storage and Offloading (FPSO) units have the advantages of their ability to provide storage and offloading capabilities which are not available in other types of floating production systems. In addition, FPSOs also provide a large deck area and substantial topsides payload capacity. They are in use in a variety of water depths and environments around the world. It is a good solution for offshore oil and gas development in fields where there is lack of an export pipeline system to shore. However due to their inherently high motions in waves, they are limited in the types of risers they can host. The Low Motion FPSO (LM-FPSO) is a novel design that is developed to maintain the advantages of the conventional FPSOs while offering significantly lower motion responses. The LM-FPSO design generally consists of a box-shape hull with large storage capacity, a free-hanging solid ballast tank (SBT) located certain distance below the hull keel, a few groups of tendons arranged to connect the SBT to the hull, a mooring system for station keeping, and a riser system. The addition of SBT to the floater results in a significant increase in heave, roll and pitch natural periods, mainly through the mass and added mass of the SBT, which significantly reduces motions in the wave frequency range. Model tests were performed at the Korea Research Institute of Ships & Ocean Engineering (KRISO) in the fall of 2016. An analytical model of the basin model (MOM) was created in Orcaflex and calibrated against the basin-model. Good agreement is achieved between global performance results from MOM's predictions and basin model measurements. The model test measurements have further verified the superior motion response of LM-FPSO. In this paper, numerical results are presented to demonstrate the comparison and correlation of the MOM results with model test measurements. The verification of the superior motion response through model test measurements is also presented in this paper.

• Journal of Veterinary Clinics
• /
• v.29 no.2
• /
• pp.141-147
• /
• 2012

This work describes the characteristics of $Malassezia$ $pachydermatis$ isolated from dog ear canals and the effect of essential oils on the growth of this organism. Sterile cotton swabs were used to collect specimens from the external ear canal and culture tests were performed to detect the population size of $Malassezia$ yeast. Using three different isolation media, included Sabouraud dextrose agar (SDA) to isolate common $M.$ $pachydermatis$, and SDA supplemented with olive oil (SDAO) and Leeming's medium (LM) to detect lipophilic yeast, $Malassezia$ spp were isolated from 14 of 18 dogs (77.8%); isolation rates were 33.3% in SDA, 72.2% in SDAO and 66.7% in LM media. All $Malassezia$ spp isolates were identified as $M.$ $pachydermatis$ according to results of PCR amplification, but gross colony morphology and SDA growth rates suggested four different subtypes. Large (LC) and medium colony (MC) types respectively describe large colony (diameter > 3 mm) and medium colony (around 2 mm) after 72 hour incubation, and small (SC) type refers to smaller colony (< 1 mm) even after 5 days incubation; lipid dependent colonies did not grow onto SDA. Large Colony type strains were isolated from 4, 11, and 11 samples, MC type strains from 2, 3 and 1 and SC type strains from 1, 2 and 1 in SDA, SDAO and LM, respectively. Lipid-dependent $M.$ $pachydermatis$ (Lipo) were isolated from 3 samples each in SDAO and LM. Anti-$M.$ $pachydermatis$ activity testing was done using disc-diffusion assays and well diffusion tests. Most essential oils inhibited the growth of $M.$ $pachydermatis$ in a range from 0.5% to 1.0% of essential oils. MIC90 and MIC50 were variable depending upon the nature of essential oils. Thyme oil was found to be highly effective in inhibiting the growth of $M.$ $pachydermatis$ in a range from 0.125% to 0.0625% while marjoram and then tea tree oil exhibited lower inhibitory capacity.