• Speech Sciences
• /
• v.5 no.1
• /
• pp.7-21
• /
• 1999

This paper proposes an improved training procedure in speech recognition based on the continuous density of the Hidden Markov Model (CDHMM). Of the three parameters (initial state distribution probability, state transition probability, output probability density function (p.d.f.) of state) governing the CDHMM model, we focus on the third parameter and propose an efficient algorithm that determines the p.d.f. of each state. It is known that the resulting CDHMM model converges to a local maximum point of parameter estimation via the iterative Expectation Maximization procedure. Specifically, we propose two independent algorithms that can be embedded in the segmental K -means training procedure by replacing relevant key steps; the adaptation of the number of mixture Gaussian p.d.f. and the initialization using the CDHMM parameters previously estimated. The proposed adaptation algorithm searches for the optimal number of mixture Gaussian humps to ensure that the p.d.f. is consistently re-estimated, enabling the model to converge toward the global maximum point. By applying an appropriate threshold value, which measures the amount of collective changes of weighted variances, the optimized number of mixture Gaussian branch is determined. The initialization algorithm essentially exploits the CDHMM parameters previously estimated and uses them as the basis for the current initial segmentation subroutine. It captures the trend of previous training history whereas the uniform segmentation decimates it. The recognition performance of the proposed adaptation procedures along with the suggested initialization is verified to be always better than that of existing training procedure using fixed number of mixture Gaussian p.d.f.

• Journal of the Korean Society of Environmental Restoration Technology
• /
• v.25 no.3
• /
• pp.47-62
• /
• 2022

The Ministry of Environment has been working to reduce the impact on biodiversity, ecosystems, and social costs caused by soil runoff from highland Agricultural fields by setting up non-point pollution source management districts. To reduce soil loss, runoff path reduction technology has been applied, but it has been less cost effective. In addition, non-point pollution sources cause environmental conflicts in downstream areas, and recently highland Agricultural fields are becoming vulnerable to climate change. The Ministry of Environment is promoting the optimal management plan in earnest to convert arable land into forests and grasslands, but since non-point pollution is not a simple environmental problem, it is necessary to approach it from the aspect of NbS(Nature-Based Solution). In this study, a scenario for applying the nature-based solution was established for three subwatersheds west of Haean-myeon, Yanggu-gun, Gangwon-do. The soil loss distribution was spatialized through GeoWEPP and the amount of soil loss was compared for the non-point pollution reduction effect of mixed forests and grasslands. When cultivated land with a slope of 20% or more and ginseng fields were restored to perennial grasslands and mixed forests, non-point pollution reduction effects of about 32% and 29.000 tons compared to the current land use were shown. Also, it was confirmed that mixed forest rather than perennial grassland is an effective nature-based solution to reduce non-point pollution.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.6
• /
• pp.677-683
• /
• 2023

The purpose of the study was to estimate the trajectories of elementary school adjustment in grades 1 through 4. For the purpose, the Korean Children's Panel data were analyzed using potential growth model and the growth mixture model. As the result, the linear model was selected as the optimal model. The four potential groups were derived as trajectories: high-level maintenance, low-level maintenance, low-level increase, and high-level decrease. In terms of group distribution, the most children were in high-level maintenance group and then low-level maintenance, low-level increase, and high-level decrease in order. Based on the findings that trajectories of elementary school adjustment changes as children growth, we suggest that schools and families need to carefully investigate and support their school adjustment in individual levels.

• Journal of the Korean GEO-environmental Society
• /
• v.19 no.5
• /
• pp.23-31
• /
• 2018

The dredging project became more important with the recent construction of off shore structures and reclamation projects. Accordingly, more exact quantitative estimation of the dredged amount should be required. The sub-sea ground information can be obtained generally by the boring investigation and the dredged amount can be estimated based on the depth or the deeper bound of a ceratin layer via 3D visualization program. During the estimation process, the input DB should be constructed with 1D elevation information from boring investigation for the spatially approximated distribution of a deeper bound of each ground layer. The input DB can be varied with the application of the borings and the approximation targets. Therefore, the 3D visualized ground profile and dredged amounts are compared on the actively dredged sites, vicinity of Saemangeum area and outer port area in Gunsan with regard to the input DB construction methods. Conclusively, the input DB based on the spatially approximated depths show higher precision results and more reasonable 3D visualized ground profiles.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.32 no.1
• /
• pp.21-30
• /
• 2022

Objectives: The purpose of this study is to suggest an optimal method by comparing the analysis methods of work environment measurement datasets including left-censored data where one or more measurements are below the limit of detection (LOD). Methods: A computer program was used to generate left-censored datasets for various combinations of censoring rate (1% to 90%) and sample size (30 to 300). For the analysis of the censored data, the simple substitution method (LOD/2), β-substitution method, maximum likelihood estimation (MLE) method, Bayesian method, and regression on order statistics (ROS)were all compared. Each method was used to estimate four parameters of the log-normal distribution: (1) geometric mean (GM), (2) geometric standard deviation (GSD), (3) 95th percentile (X95), and (4) arithmetic mean (AM) for the censored dataset. The performance of each method was evaluated using relative bias and relative root mean squared error (rMSE). Results: In the case of the largest sample size (n=300), when the censoring rate was less than 40%, the relative bias and rMSE were small for all five methods. When the censoring rate was large (70%, 90%), the simple substitution method was inappropriate because the relative bias was the largest, regardless of the sample size. When the sample size was small and the censoring rate was large, the Bayesian method, the β-substitution method, and the MLE method showed the smallest relative bias. Conclusions: The accuracy and precision of all methods tended to increase as the sample size was larger and the censoring rate was smaller. The simple substitution method was inappropriate when the censoring rate was high, and the β-substitution method, MLE method, and Bayesian method can be widely applied.

• Journal of Bio-Environment Control
• /
• v.29 no.2
• /
• pp.161-170
• /
• 2020

There is systematic spatial variations in environmental properties due to sensitive reaction to external conditions at plastic greenhouse occupied 99.2% of domestic agricultural facilities. In order to construct 3 dimensional distribution of temperature, relative humidity, CO₂ and illuminance, measurement matrix as 3 by 3 by 5 in direction of width, height and length, respectively, dividing indoor space of greenhouse was designed and tested at experimental site. Linear regression analysis was conducted to evaluate optimal estimation method in terms with horizontal and vertical variations. Even though sole measurement point for temperature and relative humidity could be feasible to assess indoor condition, multiple measurement matrix is inevitably required to improve spatial precision at certain time domain such as period of sunrise and sunset. In case with CO₂, multiple measurement matrix could not successfully improve the spatial predictability during a whole experimental period. In case with illuminance, prediction performance was getting smaller after a time period of sunrise due to systematic interference such as indoor structure. Thus, multiple sensing methodology was proposed in direction of length at higher height than growing bed, which could compensate estimation error in spatial domain. Appropriate measurement matrix could be constructed considering the transition of stability in indoor environmental properties due to external variations. As a result, optimal measurement matrix should be carefully designed considering flexibility of construction relevant with the type of property, indoor structure, the purpose of crop and the period of growth. For an instance, partial cooling and heating system to save a consumption of energy supplement could be successfully accomplished by the deployment of multiple measurement matrix.

• Journal of Korea Water Resources Association
• /
• v.43 no.12
• /
• pp.1011-1027
• /
• 2010

In hydrologic modeling, prediction uncertainty generally stems from various uncertainty sources associated with model structure, data, and parameters, etc. This study aims to assess the parameter uncertainty effect on hydrologic prediction results. For this objective, a distributed rainfall-sediment yield-runoff model, which consists of rainfall-runoff module for simulation of surface and subsurface flows and sediment yield module based on unit stream power theory, was applied to the mesoscale mountainous area (Cheoncheon catchment; 289.9 $km^2$). For parameter uncertainty evaluation, the model was calibrated by a multi-objective optimization algorithm (MOSCEM) with two different objective functions (RMSE and HMLE) and Pareto optimal solutions of each case were then estimated. In Case I, the rainfall-runoff module was calibrated to investigate the effect of parameter uncertainty on hydrograph reproduction whereas in Case II, sediment yield module was calibrated to show the propagation of parameter uncertainty into sedigraph estimation. Additionally, in Case III, all parameters of both modules were simultaneously calibrated in order to take account of prediction uncertainty in rainfall-sediment yield-runoff modeling. The results showed that hydrograph prediction uncertainty of Case I was observed over the low-flow periods while the sedigraph of high-flow periods was sensitive to uncertainty of the sediment yield module parameters in Case II. In Case III, prediction uncertainty ranges of both hydrograph and sedigraph were larger than the other cases. Furthermore, prediction uncertainty in terms of spatial distribution of erosion and deposition drastically varied with the applied model parameters for all cases.

• Korean Journal of Remote Sensing
• /
• v.36 no.2_1
• /
• pp.179-197
• /
• 2020

Topographic normalization reduces the terrain effects on reflectance by adjusting the brightness values of the image pixels to be equal if the pixels cover the same land-cover. Topographic effects are induced by the imaging conditions and tend to be large in high mountainousregions. Therefore, image analysis on mountainous terrain such as estimation of wildfire damage assessment requires appropriate topographic normalization techniques to yield accurate image processing results. However, most of the previous studies focused on the evaluation of topographic normalization on satellite images with moderate-low spatial resolution. Thus, the alleviation of topographic effects on multi-temporal high-resolution images was not dealt enough. In this study, the evaluation of terrain normalization was performed for each band to select the optimal technical combinations for rapid and accurate wildfire damage assessment using PlanetScope images. PlanetScope has considerable potential in the disaster management field as it satisfies the rapid image acquisition by providing the 3 m resolution daily image with global coverage. For comparison of topographic normalization techniques, seven widely used methods were employed on both pre-fire and post-fire images. The analysis on bi-temporal images suggests the optimal combination of techniques which can be applied on images with different land-cover composition. Then, the vegetation index was calculated from the images after the topographic normalization with the proposed method. The wildfire damage detection results were obtained by thresholding the index and showed improvementsin detection accuracy for both object-based and pixel-based image analysis. In addition, the burn severity map was constructed to verify the effects oftopographic correction on a continuous distribution of brightness values.

• Journal of Korean Society of Transportation
• /
• v.19 no.5
• /
• pp.49-59
• /
• 2001

In the transportation planning process, origin and destination(O-D) trip matrix is one of the most important elements. There have been developments and applications of the methodology to adjust old matrices using link traffic counts. Commonly, the accuracy of an adjusted O-D matrix depends very much on the reliability of the input data such as the numbers and locations of traffic counting points in the road network. In the real application of the methodology, decisions on the numbers and locations of traffic counting points are one of the difficult problems, because usually as networks become bigger, the numbers of traffic counting points are required more. Therefore, this paper investigates these issues as an experiment using a nationwide network in Korea. We have compared and contrasted the set of link flows assigned by the old and the adjusted O-D matrices with the set of observed link flows. It has been analyzed by increasing the number of the traffic counting points on the experimental road network. As a result of these analyses, we can see an optimal set of the number of counting links through statistical analysis, which are approximately ten percentages of the total link numbers. In addition, the results show that the discrepancies between the old and the adjusted matrices in terms of the trip length frequency distributions and the assigned and the counted link flows are minimized using the optimal set of the counted links.

• Journal of Korea Water Resources Association
• /
• v.38 no.7 s.156
• /
• pp.545-554
• /
• 2005

The detailed interdisciplinary surveys were conducted on the pattern of habitat use of dominant fishes during the spawning and adult stage. The hydraulic parameters of the depth and velocity, discharge, substrate cover streams, and distribution of fish in the Yeonggang, WiCheon, HoeCheon, GeochangwiCheon, CheongdoCheon, DanjangCheon (the Nakdong River Basin) were measured. The Habitat Suitability Criteria was developed for the two fish species (Zacco Platypus and Zacco Temmincki) and life stages(spawning and adult), habitat conditions (depth, velocity and covet). The Physical Habitat Simulation of the Instream Flow Incremental Methodology was applied to calculate for optimal flow and the ecological flow recommendation was proposed by choosing the largest one in the optimal flow. The ecological flow recommendation was $5.0\;m^3/s{\sim}10.0\;m^3/s$ (e.g., $6.5\;m^3/s$ in the NaeseongCheon). Also, the ecological flow recommendations were compared with the existing ecological flow and flow duration analysis.