• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.335-341
• /
• 2022

In this paper, we adopt a transformer that shows remarkable performance in natural language processing as an acoustic model of hybrid speech recognition. The transformer acoustic model uses attention structures to process sequential data and shows high performance with low computational cost. This paper proposes a method to improve the performance of transformer AM by applying each of the four algorithms of sequence discriminative training, a weighted finite-state transducer (wFST)-based learning used in the existing DNN-HMM model. In addition, compared to the Cross Entropy (CE) learning method, sequence discriminative method shows 5 % of the relative Word Error Rate (WER).

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.49 no.1
• /
• pp.18-28
• /
• 2013

It was compared the estimated parameters by the surplus production from three different models, i.e., three types (Schaefer, Gulland, and Schnute) of the traditional surplus production models, a stock production model incorporating covariates (ASPIC) model and a maximum entropy (ME) model. We also evaluated the performance of models in the estimation of their parameters. The maximum sustainable yield (MSY) of small yellow croaker (Pseudosciaena polyactis) in Korean waters ranged from 35,061 metric tons (mt) by Gulland model to 44,844mt by ME model, and fishing effort at MSY ($f_{MSY}$) ranged from 262,188hauls by Schnute model to 355,200hauls by ME model. The lowest root mean square error (RMSE) for small yellow croaker was obtained from the Gulland surplus production model, while the highest RMSE was from Schnute model. However, the highest coefficient of determination ($R^2$) was from the ME model, but the ASPIC model yielded the lowest coefficient. On the other hand, the MSY of Kapenta (Limnothrissa miodon) ranged from 16,880 mt by ASPIC model to 25,373mt by ME model, and $f_{MSY}$, from 94,580hauls by ASPIC model to 225,490hauls by Schnute model. In this case, both the lowest root mean square error (RMSE) and the highest coefficient of determination ($R^2$) were obtained from the ME model, which showed relatively better fits of data to the model, indicating that the ME model is statistically more stable and robust than other models. Moreover, the ME model could provide additional ecologically useful parameters such as, biomass at MSY ($B_{MSY}$), carrying capacity of the population (K), catchability coefficient (q) and the intrinsic rate of population growth (r).

• The KIPS Transactions:PartB
• /
• v.15B no.1
• /
• pp.61-68
• /
• 2008

This paper presents a spam filter system using co-training with spamminess features and URL features based on the maximum entropy model. Spamminess features are the emphasizing patterns or abnormal patterns in spam messages used by spammers to express their intention and to avoid being filtered by the spam filter system. Since spammers use URLs to give the details and make a change to the URL format not to be filtered by the black list, normal and abnormal URLs can be key features to detect the spam messages. Co-training with spamminess features and URL features uses two different features which are independent each other in training. The filter system can learn information from them independently. Experiment results on TREC spam test collection shows that the proposed approach achieves 9.1% improvement and 6.9% improvement in accuracy compared to the base system and bogo filter system, respectively. The result analysis shows that the proposed spamminess features and URL features are helpful. And an experiment result of the co-training shows that two feature sets are useful since the number of training documents are reduced while the accuracy is closed to the batch learning.

• Journal of Korean Society of Transportation
• /
• v.1 no.1
• /
• pp.43-47
• /
• 1983

Within the conventional transportation planning process, "trip distribution" has a significant role to play. The most widely applied trip distribution model is the gravity model, for which Wilson provided the theoretical basis in 1967. The concept of the gravity model, however, still remains ambiguous if we analyze the "trip distribution" with a disaggregate data set. Thus, this paper hypothesizes that the gravity technique is still valid even with the disaggregate data set, by proving that the estimated coefficients of the gravity model, which is derived under the principle of entropy maximization, are identical with those of the multinomial logit model, which is derived under the principle of individual utility maximization.tility maximization.

• Korean Chemical Engineering Research
• /
• v.47 no.4
• /
• pp.481-487
• /
• 2009

In this of study, effects of input air velocity(0.05~0.22 m/sec) and molten carbonate salt temperature ($870{\sim}970^{\circ}C$) on flow regime transition have been studied by adopting a drift-flux model of air holdup and a stochastic analysis of differential pressure fluctuations in an air-molten sodium carbonate salt two-phase system(molten salt oxidation process). Air holdup where the flow regime transition begins was determined by air holdup-drift flux plot. The air holdup value which the flow regime transition begins was increased with increasing molten carbonate salt temperature due to the decrease of viscosity and surface tension of molten carbonate salt. To characterize the flow regime transition more quantitatively, differential pressure fluctuation signals have been analyzed by adopting the stochastic method such as phase space portraits and Kolmogorov entropy, The Kolmogorov entropy decreased with an increasing of molten carbonate salt temperature but increased gradually with an increase in an air velocity, however, it exhibited different tendency with the flow regime and the air velocity value which flow regime transition begins was same to the results of drift-flux analysis.

• The Korean Journal of Applied Statistics
• /
• v.19 no.3
• /
• pp.521-534
• /
• 2006

This paper considers a panel regression model with ill-posed data and proposes the generalized maximum entropy(GME) estimator of the unknown parameters. These are natural extensions from the biometries, statistics and econometrics literature. The performance of this estimator is investigated by using of Monte Carlo experiments. The results indicate that the GME method performs the best in estimating the unknown parameters.

• Geomechanics and Engineering
• /
• v.12 no.1
• /
• pp.35-52
• /
• 2017

In order to evaluate the grouting effects of water-rich fault in tunnels systematically, a feasible and scientific method is introduced based on the extension theory. First, eight main influencing factors are chosen as evaluation indexes by analyzing the changes of permeability, mechanical properties and deformation of surrounding rocks. The model of evaluating grouting effects based on the extension theory is established following this. According to four quality grades of grouting effects, normalization of evaluation indexes is carried out, aiming to meet the requirement of extension theory on data format. The index weight is allocated by adopting the entropy method. Finally, the model is applied to the grouting effects evaluation in water-rich fault F4-4 of Qingdao Jiaozhou Bay Subsea Tunnel, China. The evaluation results are in good agreement with the test results on the site, which shows that the evaluation model is feasible in this field, providing a powerful tool for systematically evaluating the grouting effects of water-rich fault in tunnels.

• Communications for Statistical Applications and Methods
• /
• v.16 no.2
• /
• pp.265-275
• /
• 2009

This paper considers a linear regression model with a spatial autoregressive disturbance with ill-posed data and proposes the generalized maximum entropy(GME) estimator of regression coefficients. The performance of this estimator is investigated via Monte Carlo experiments. The results show that the GME estimator provides efficient and robust estimate for the unknown parameter.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.3
• /
• pp.626-634
• /
• 2012

It is required that system can seamlessly identify and manage change history and comprehensive assessment of several types of data as well as individual information of feeding and water environment for scientific and systematic management of fish farming environment and fish farmer. In this study, we implemented the system which can present and simulate current status of water quality and feeding based on th historical data of them, and check changes of state step by step using visual C++. In addition, we proposed the entropy model which can be comprehensive analysis about water quality and feed status information based on knowledge of fisheries. It can be the foundation to create high-level environment model reflecting the more diverse fisheries knowledge such as disease.

• Journal of Forest and Environmental Science
• /
• v.32 no.3
• /
• pp.253-261
• /
• 2016

Xanthoceras sorbifolia seed coat (XSSC) is a processing residue of the bioenergy crop. This work aimed to evaluate the applicability of using the steam explosion to modify the residue for dye biosorption from aqueous solutions by using methylene blue as a model cationic dye. Equilibrium, kinetic and thermodynamic parameters for the biosorption of methylene blue on the steam-exploded XSSC (SE-XSSC) were evaluated. The kinetic data followed the pseudo-second-order model, and the rate-limiting step was the chemical adsorption. Intraparticle diffusion was one of the rate-controlling factors. The equilibrium data agreed well with the Langmuir isotherm, and the biosorption was favorable. The steam-explosion pretreatment strongly affected the biosorption in some respects. It reduced the adsorption rate constant and the initial sorption rate of the pseudo-second-order model. It enhanced the adsorption capacity of methylene blue at higher temperatures while reduced the capacity at lower ones. It changed the biosorption from an exothermic process driven by both the enthalpy and the entropy to an endothermic one driven by entropy only. It increased the surface area and decreased the pH point of zero charge of the biomass. Compared with the native XSSC, SE-XSSC is preferable to MB biosorption from warmer dye effluents.