• Journal of KIISE:Software and Applications
• /
• v.29 no.12
• /
• pp.966-978
• /
• 2002

There are two important problems in improving text classification systems based on machine learning approach. The first one, called "selection problem", is how to select a minimum number of informative documents from a given document collection. The second one, called "composition problem", is how to reorganize selected training documents so that they can fit an adopted learning method. The former problem is addressed in "active learning" algorithms, and the latter is discussed in "boosting" algorithms. This paper proposes a new learning method, called AdaBUS, which proactively solves the above problems in the context of Naive Bayes classification systems. The proposed method constructs more accurate classification hypothesis by increasing the valiance in "weak" hypotheses that determine the final classification hypothesis. Consequently, the proposed algorithm yields perturbation effect makes the boosting algorithm work properly. Through the empirical experiment using the Routers-21578 document collection, we show that the AdaBUS algorithm more significantly improves the Naive Bayes-based classification system than other conventional learning methodson system than other conventional learning methods

• Korean Journal of Remote Sensing
• /
• v.34 no.2_1
• /
• pp.203-211
• /
• 2018

Dam discharge is carried out for the management of rivers and area around rivers due to rainy season or drought. Dam discharge should be based on an accurate understanding of the flow rate that can be accommodated in the river. Therefore, understanding the allowable storage capacity of river is an important factor in the management of the environment around the river. However, the methods using water level meters and images, which are currently used to determine the allowable flow rate of rivers, show limitations in terms of accuracy and efficiency. In order to solve these problems, this paper proposes a method to automatically calculate the allowable storage capacity of river based on the images taken by drone. In the first step, we create a 3D model of the river by using the drone images. This generation process consists of tiepoint extraction, image orientation, and image matching. In the second step, the allowable storage capacity is calculated by cross section analysis of the river using the generated river 3D model and the road and river layers in the target area. In this step, we determine the maximum water level of the river, extract the cross-sectional profile along the river, and use the 3D model to calculate the allowable storage capacity for the area. To prove our method, we used Bukhan river's data and as a result, the allowable storage volume was automatically extracted. It is expected that the proposed method will be useful for real - time management of rivers and surrounding areas and 3D models using drone.

• The Korean Journal of Applied Statistics
• /
• v.28 no.6
• /
• pp.1171-1180
• /
• 2015

Multidimensional scaling (MDS) is a graphical technique of multivariate analysis to display dissimilarities among individuals into low-dimensional space. We often have two kinds of MDS which are metric MDS and non-metric MDS. Metric MDS can be applied to quantitative data; however, we need additional information about variables because it only shows relationships among individuals. Gower (1992) proposed a method that can represent variable information using trajectories of the pseudo-points for quantitative variables on the metric MDS space. We will call his method a 'replacement method'. However, the trajectory can not be represented even though metric MDS can be applied to binary data when we apply his method to binary data. Therefore, we propose a method to represent information of binary variables using pseudo-points called a 'partition method'. The proposed method partitions pseudo-points, accounting both the rate of zeroes and ones. Our metric MDS using the proposed partition method can show the relationship between individuals and variables for binary data.

• Korean Journal of Soil Science and Fertilizer
• /
• v.45 no.6
• /
• pp.943-948
• /
• 2012

This study was conducted to estimate the effect of nitrogen reduction by applying ryegrass and to determine the relationship between yield and nitrate concentration of soil solution for cucumber cultivation in plastic film house. Nitrogen levels with recovery of ryegrass ($42.3Mg\;ha^{-1}$) was 0, 50, 75, 100 % of 199 kg N $ha^{-1}$ as N recommendation by determining soil EC value. Yield and nitrate concentration in soil solution was investigated during cucumber cultivation. Yields of N treatments applied ryegrass showed 64.3, 70.9, 70.3, and $76.5Mg\;ha^{-1}$, respectively, it could reduce about 25-50% of nitrogen application compared to yield ($68Mg\;ha^{-1}$) of NPK plot applied 199 kg N $ha^{-1}$. Nitrate concentration in soil solution was average 26.0, 30.1, 41.4, $58.5mg\;L^{-1}$ during cucumber cultivation and was related between yield and average nitrate concentration of soil solution following as; $Y=49.3+0.63X+0.0034X^2$ ($R^2=0.778^{**}$). However, it needs to conduct extra-experiment due to high variation of nitrate concentration during cultivation periods.

• Journal of Korean Society of Transportation
• /
• v.24 no.6 s.92
• /
• pp.33-43
• /
• 2006

This research is to the selection of optimal probability distribution as well as the estimation for design hourly factor in consideration of traffic characteristic, such as road function, lane number and AADT. To accomplish the objectives, we are applied to various probability distribution using traffic data that observed at permanent traffic count points in 2005. The parameters or the selected 14 probability distribution were estimated based on the method of maximum likelihood and the validity condition of the estimated parameter The goodness-of-fit test, such as chi-square test. was performed as well as the estimation of design hourly factor. As a result, An appropriate distributions of each case were selected : Pearson V for two lane of rural roads, LogLogistic for the four lane of rural roads, LogLogistic for the urban roads, Extreme value for recreation roads. And optimal K factor are as following : $0.1{\sim}0.2 $ for two lane of rural roads, $0.09{\sim}0.14$ for the four lane of rural roads. $0.07{\sim}0.13$ for the urban roads, $0.1{\sim}0.2$ for recreation roads.

• Journal of the Korean Vacuum Society
• /
• v.20 no.1
• /
• pp.50-56
• /
• 2011

In this study, the effects of ZnO buffer layer thickness and annealing temperature on the heterojunction diode, ZnO/b-ZnO/p-Si(111), were reported. The effects of those on the structural and electrical properties of zinc oxide (ZnO) films on ZnO buffered p-Si (111) substrate were also studied. Structural properties of ZnO thin films were studied by X-ray diffraction and I-V characteristics were measured by a semiconductor parameter analyzer. ZnO thin films with 70 nm thick buffer layer and annealing temperature of $700^{\circ}C$ showed the best c-axis preferred orientation. The best electrical property was found at the condition of buffer layer annealing temperature of $700^{\circ}C$ and 50nm thick ZnO buffer layer (resistivity: $2.58{\times}10^{-4}[{\Omega}-cm]$, carrier concentration: $1.16{\times}1020[cm^{-3}]$). The I-V characteristics for ZnO/b-ZnO/p-Si(111) heterojunction diode were improved with increasing buffer layer thickness at buffer layer annealing temperature of $700^{\circ}C$.

• Economic and Environmental Geology
• /
• v.50 no.2
• /
• pp.105-116
• /
• 2017

The purpose of this study is to provide the information on genesis of obsidian occurring in the southwestern part of Ulleung Island, Korea, and to discuss its implications for volcanic activity through volcanological and mineralogical properties of obsidian. Obsidian occurs locally at the lower part of the Gombawi welded tuff, showing various complex textures and flow banding. Though obsidian is mostly homogeneous, it is closely associated with alkali feldspar phenocrysts, reddish tuff, and greyish trachyte fragments. The obsidian occurs as wavy, lenticular blocks or lamination composed of fragments. Cooling fractures developed on obsidian glass are characterized by perlitic cracks, orbicular or spherical cracks, indicating that obsidian rapidly quenched to form an amorphous silica-rich phase. It is evident that hydration took place preferentially at the outer rim relative to the core of obsidian, forming alteration rinds. The glassy matrix of obsidian includes euhedral alkali feldspars, diopside, biotite, ilmenite, and iron oxides. Microlites in glassy obsidian are composed mainly of alkali feldspars and ilmenite. Quantitative analysis by EPMA on the obsidian glass part shows trachytic composition with high iron content of 3 wt.%. Accordingly, obsidian formed with complex textures under a rapid cooling condition on surface ground, with slight rheomorphism. Such results might be induced by collapse of lava dome or caldera, which produced the block-and-ash flow deposit and the transportation into valley while keeping high temperatures.

• Microbiology and Biotechnology Letters
• /
• v.39 no.4
• /
• pp.324-329
• /
• 2011

Eight endophytic fungal strains were isolated from the roots of Calystegia soldanella from the western coast of South Korea. The culture filtrate of the eight endophytic fungi were applied to waito-c rice seedlings in order to verify potential plant growth promotion activities. The results of bioassay indicated that the Cs-9-7 fungal strain possessed the highest plant growth promotion activity. Fungal culture filtrates were analyzed to verify secondary metabolites using gas chromatography and mass spectroscopy with selected ion monitoring (GC/MS-SIM). The culture filtrate of the Cs-9-7 fungal strain was confirmed to contain gibberellins GA3 (1.229 ng/mL), GA4 (3.535 ng/mL), GA7 (1.408 ng/mL) and GA12 (0.378 ng/mL). Polymerase chain reactions (PCR) were performed so as to determine the internal transcribed spacer (ITS) regions for the identification of isolated strains with universal primers ITS-1 and ITS-4. The Cs-9-7 fungal strain, isolated from the root of C. soldanella, has been named Aspergillus tubingensis Cs-9-7.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.9 no.3
• /
• pp.195-202
• /
• 2007

The normalized difference in incident solar energy between a target surface and a level surface (overheating index, OHI) is useful in eliminating estimation error of site-specific maximum temperature in complex terrain. Due to the complexity in its calculation, however, an empirical proxy variable called northern exposure index (NEI) which combines slope and aspect has been used to estimate OHI based on empirical relationships between the two. An experiment with real-world landscape and temperature data was carried out to evaluate performance of the NEI - derived OHI (N-OHI) in reduction of spatial interpolation error for daily maximum temperature compared with that by the original OHI. We collected daily maximum temperature data from 7 sites in a mountainous watershed with a $149 km^2$ area and a 795m elevation range ($651{\sim}1,445m$) in Pyongchang, Kangwon province. Northern exposure index was calculated for the entire 166,050 grid cells constituting the watershed based on a 30-m digital elevation model. Daily OHI was calculated for the same watershed ana regressed to the variation of NEI. The regression equations were used to estimate N-OHI for 15th of each month. Deviations in daily maximum temperature at 7 sites from those measured at the nearby synoptic station were calculated from June 2006 to February 2007 and regressed to the N-OHI. The same procedure was repeated with the original OHI values. The ratio sum of square errors contributable by the N-OHI were 0.46 (winter), 0.24 (fall), and 0.01 (summer), while those by the original OHI were 0.52, 0.37 and 0.15, respectively.

• Journal of Animal Science and Technology
• /
• v.45 no.4
• /
• pp.641-648
• /
• 2003

In a batch mode treatment process, which electrolyzes the wastewater after derivation of N-P crystal formation and recovery, the characteristics of pollutant removal induced with the changes of loading rate and hydraulic retention time were studied. $MgCl_2$ was used as Mg source for the formation of struvite and the molar ratio of $MgCl_2$ to $PO_4^{3-}$ in influent was 1.3. When analyzing the average treatment efficiencies and removal characteristics obtained from four separate operations (Run I, II, III, IV), removal efficiencies of PO43- was not function of its loading rate. Under a condition of sufficient aeration and Mg source provided, over 88% of $PO_4^{3-}$ was eliminated by the formation of MAP without any pH adjustment, in spite of loading rate variation. An optimum-loading rate of NH4-N to achieve high removal efficiency was approximately $100g/m^3/d$. Below that loading rate, the removal of NH4-N was proportional to the loading rate into the system, and hence stable and high removal efficiency, over 90%, was achieved. However, when the loading rate increased over that rate, removal efficiency began to drop and fluctuate. Removal efficiency of TOCs was dependant upon the hydraulic retention time ($r^2$=0.97), not upon the loading rate. Stable and high color removal (94%) was obtained with 2 days of HRT in electrolysis reactor.