• The Korean Journal of Applied Statistics
• /
• v.30 no.6
• /
• pp.867-876
• /
• 2017

Datasets with small n and large p are often found in various fields and the analysis of the datasets is still a challenge in statistics. Discriminant analysis models for such datasets were recently developed in classification problems. One approach of those models tries to detect dimensions that distinguish between groups well and the number of the detected dimensions is typically smaller than p. In such models, the number of dimensions is important because the prediction and visualization of data and can be usually determined by the K-fold cross-validation (CV). However, in sparse data scenarios, the CV is not reliable for determining the optimal number of dimensions since there can be only a few observations for each fold. Thus, we propose a method to determine the number of dimensions using a measure based on the standardized distance between the mean values of each group in the reduced dimensions. The proposed method is verified through simulations.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.67 no.4
• /
• pp.183-190
• /
• 2018

Predicting accurate electricity prices is an important task in the electricity trading market. To address the electricity price forecasting problem, various approaches have been proposed so far and it is known that linear regression-based approaches are the best. However, the use of such linear regression-based methods is limited due to low accuracy and performance. In traditional linear regression methods, it is not practical to find a nonlinear regression model that explains the training data well. If the training data is complex (i.e., small-sized individual data and large-sized features), it is difficult to find the polynomial function with n terms as the model that fits to the training data. On the other hand, as a linear regression model approximating a nonlinear regression model is used, the accuracy of the model drops considerably because it does not accurately reflect the characteristics of the training data. To cope with this problem, we propose a new electricity price forecasting method that divides the entire dataset to multiple split datasets and find the best linear regression models, each of which is the optimal model in each dataset. Meanwhile, to improve the performance of the proposed method, we modify the proposed localized linear regression method in the map and reduce way that is a framework for parallel processing data stored in a Hadoop distributed file system. Our experimental results show that the proposed model outperforms the existing linear regression model. Specifically, the accuracy of the proposed method is improved by 45% and the performance is faster 5 times than the existing linear regression-based model.

• Neonatal Medicine
• /
• v.18 no.1
• /
• pp.49-58
• /
• 2011

Purpose: The purpose of this study was to evaluate the relationship between cord blood adiponectin and insulin-like growth factor (IGF)-I and their effect on fetal growth and insulin resistance in mothers with gestational diabetes mellitus (GDM). Methods: Cord blood adiponectin and IGF-I were compared between mothers with GDM (GDM group, N=53) and controls (non-GDM group, N=101). Neonates were classified into three groups of small for gestational age (SGA, N=26), appropriate for gestational age (AGA, N=97), and large for gestational age (LGA, N=31) by birth weight. The association between cord adiponectin and IGF-I levels was evaluated in relation to maternal and neonatal clinical data. Results: Cord adiponectin was lower in the GDM group than in the non-GDM group (P<0.001). There was no significant difference in cord adiponectin among the SGA, AGA, and LGA groups in the GDM group (P=0.228). The cord adiponectin of AGA in the GDM group was significantly lower than that in the non-GDM group (P<0.001). The most powerful predictor affecting cord adiponectin was the result of maternal 75 g oral glucose tolerance test. The cord IGF-I values between the GDM group and the non-GDM group were not different (P=0.834). Neonates with the heavier birth weight had the higher cord IGF-I levels. The most powerful predictor affecting cord IGF-I was birth weight and the next was maternal parity. Conclusion: Both cord blood adiponectin and IGF-I were associated with fetal growth, but IGF-I was a more general and direct factor affecting fetal body size, and adiponectin seemed to have more association with insulin sensitivity than growth.

• Asian-Australasian Journal of Animal Sciences
• /
• v.19 no.5
• /
• pp.661-671
• /
• 2006

Steers (n = 335) of known genetic backgrounds from four fundamentally different growth types were subjected to two production systems to study the main effects and possible interactive effects on carcass composition. Growth types were animals with genetic potential for large mature weight (LL), intermediate mature weight-late maturing (IL), intermediate mature weight-early maturing (IE), and small mature weight-early maturing (SE). Each year, in a nine year study, calves of each growth type were weaned and five steers of each growth type were developed on pasture or feedlot and harvested at approximately 20 and 14 mo of age, respectively. Data recorded were chilled carcass weight and percentages of forequarter, foreshank, chuck, rib, plate, brisket, hindquarter, round, rump, shortloin, sirloin, flank, lean, fat, bone, and retail cuts. The growth $type{\times}production$ system interaction was an important source of variation in chilled carcass weight (p = 0.0395) and percentage retail cuts (p = 0.001), lean (p = 0.001), fat (p = 0.001), rump (p = 0.0454), shortloin (p = 0.0487), and flank (p = 0.001). The ranking of the growth $type{\times}production$ system means for percentage lean was LL-pasture>IL-pasture = IE-pasture = SE-pasture>LL-feedlot, IL-feedlot>IE-feedlot = SE-feedlot. The growth $type{\times}production$ system interaction was non-significant (p>0.05) for forequarter, foreshank, chuck, rib, plate, brisket, hindquarter, round and bone. Growth types of IE and SE yielded greater (p<0.05) mean forequarter than did growth types of IL and LL ($51.6{\pm}0.3$ and $51.5{\pm}0.3$ vs. $51.1{\pm}0.3$ and $50.8{\pm}0.3%$). Mean bone was highest (p<0.05) for the LL growth type and lowest (p<0.05) for the SE growth type ($19.5{\pm}0.5$ vs. $16.8{\pm}0.5%$). Mean bone was greater (p<0.05) for the pastured steers than for the feedlot steers ($21.8{\pm}0.8$ vs. $14.5{\pm}0.6%$). These data indicate that growth type responded differently in the two production systems and that these results should be helpful in the match of genetics to production resources.

• Tuberculosis and Respiratory Diseases
• /
• v.43 no.6
• /
• pp.945-953
• /
• 1996

Background : Many acute and chronic lung diseases including pneumoconiosis are characterized by the presence of increased numbers of activated macrophages. These macrophages generate several inflammatory cell chemoattractants, by which neutrophil migrate from vascular compartment to the alveolar space. Recruited neutrophils secrete toxic oxygen radicals or proteolytic enzymes and induce inflammatory response. Continuing inflammatory response results in alteration of the pulmonary structure and irreversible fibrosis. Recently, a polypeptide with specific neutrophil chemotactic activity, interleukin-8(IL-8), has been cloned and isolated from a number of cells including : monocytes, macrophages and fibroblasts. IL-1 and/or TNF-${\alpha}$ preceded for the synthesis of IL-8, and we already observed high level of IL-1 and TNF-${\alpha}$ in the pneumoconioses. So we hypothesized that IL-8 may be a central role in the pathogenesis of pneumoconiosis. In order to evaluate the clinical utility of IL-8 as a biomarker in the early diagnosis of pneumoconiosis, we investigated the increase of IL-8 in the pneumoconiotic patient and the correlation between IL-8 level and progression of pneumoconiosis. Method : We measured IL-8 in the serum of 48 patients with pneumoconiosis and 16 persons without dust exposure history as a control group. Pneumoconiotic cases were divided into 3 groups according to ILO Classification : suspicious group(n=16), small opacity group(n=16) and large opacity group(n=16). IL-8 was measured by a sandwich enzytne immunoassay technique. All data were expressed as the $mean{\pm}standard$ deviation. Results: 1) The mean value of age was higher in the small opacity and large opacity group than comparison group, but smoking history was even. Duration of dust exposure was not different among 3 pneumoconiosis groups. 2) IL-8 level was $70.50{\pm}53.63pg/m{\ell}$ in the suspicious group, $107.50{\pm}45.88pg/m{\ell}$ in the small opacity group, $132.50{\pm}73.47pg/m{\ell}$ in the large opacity group and $17.85{\pm}33.85pg/m{\ell}$ in the comparison group. IL-8 concentration in all pneumoconiosis group was significant higher than that in the comparison group(p<0.001). 3) IL-8 level tended to increase with the progression of pneumoconiosis. Multiple comparison test using Anova/Scheffe analysis showed a significant difference between suspicious group and large opacity group(p<0.05). 4) The level of IL-8 was correlated with the progression of pneumoconiosis(r=0.4199, p<0.05). Conclusion : IL-8 is thought to be a good biomarker for the early diagnosis of pneumoconiosis.

• Asian-Australasian Journal of Animal Sciences
• /
• v.25 no.11
• /
• pp.1634-1640
• /
• 2012

The objective of this study was to compare pork characteristics and to determine consumer acceptability of pork chops from antibiotic free Yorkshire crossbreds sired by Berkshire (BY), Large Black (LBY), Tamworth (TY) or Yorkshire (YY) boars and reared in hoop houses. The experiments were conducted at the North Carolina Agricultural and Technical State University (NCA&TSU) Farm in Greensboro, NC and the Cherry Research Station Center for Environmental Farming Systems (CEFS) Alternative Swine Unit in Goldsboro, NC (source of antibiotic free Yorkshire sows used at both places). Twenty-four sows were artificially inseminated at each location in each of three trials. Litters were weaned at 4 wks old, and reared within deep-bedded outdoor hoop houses. To compare pork characteristics, 104 randomly selected animals were harvested at a USDA-inspected abattoir at approximately 200 d of age. Variables measured included pH, color score, $L^*$, $a^*$, $b^*$, marbling score, drip loss, hot carcass weight, backfat thickness (BF), loin muscle area (LMA), and slice shear force. Sensory panel tests were also conducted at two time periods. The data was analyzed with GLM in SAS 9.01 including location, trial, and sire breed as fixed effects. Backfat thickness, LMA, color score and $a^*$ were different among breeding groups (p<0.05). The LBY pigs had thicker backfat and smaller LMA than the other breed types. The TY and YY had less backfat than all other breed groups. Color score was lower for YY than BY and LBY but intermediate for TY. The $a^*$ was lower for TY than other breeds except LBY which was intermediate. For one sensory panel test, YY pork was more preferred overall as well as for juiciness and texture compared to BY and LBY (p<0.05), but no impact of breed type was noted for the other test, with values similar for BY, LBY, TY and YY pork. This information may help small farmers make decisions about breed types to use for outdoor production.

• Journal of Korean Society on Water Environment
• /
• v.29 no.4
• /
• pp.514-522
• /
• 2013

The major 24 tributaries in Nam-River mid-watershed were monitored for discharge and water quality in order to understand the characteristics of the watershed and to select the tributary catchment for improving water quality. According to the analytical results of discharge and water quality monitoring data of 24 tributaries, the mean value of discharge below $0.1m^3/s$ was 62.5% among the monitored tributaries and it mostly exceeded the water quality standards of Nam-river mid-watershed ($BOD_5$ = 3 mg/L, T-P = 0.1 mg/L over). According to the stream grouping method and the water quality delivery load density ($kg/day/km^2$) based on the results of tributary discharge and water quality monitoring, the tributary watersheds for improving the water quality were selected. In the Nam-River mid-watershed, tributaries in the GaJwaCheon, HaChonCheon catchment (Group D, $BOD_5$ = 3 mg/L over) and in the UirYeongCheon, SeokGyoCheon catchment (Group A, T-P = 0.1 mg/L over), which have a small flow (and/or large flow) and a high concentrations of water pollutants. The various water quality improving scheme for tributaries, in accordance with the reduction of potential point source pollution by living sewage and livestock wastewater, should be established and implemented.

• Economic and Environmental Geology
• /
• v.36 no.3
• /
• pp.159-170
• /
• 2003

Heavy metals are extracted from Chonju stream sediment, roadside soils and sediments along Honam expressway, soils and tailings from mining area using three different methods (partial extraction in Standard Method, partial extraction method with maintaining 0.1 N of extraction solution and Sequential Extraction Method). In samples having buffer capacity against acid, pH 1 (0.1 N HCl) of extraction solution can not be maintained and pH of extraction solution increases up to 8.0 when partial extraction in Standard Method is used. The averages and ranges of HPE(heavy metals extracted using partial extraction in Standard Method)/HPEM(heavy metals extracted using partial extraction method with maintaining 0.1 N of extraction solution) values are 0.479 and 0.145~0.929 for Cd, 0.534 and 0.078~0.928 for Zn, 0.432 and 0.041~0.992 for Mn, 0.359 and 0.011~0.874 for Cu, 0.150 and 0.018~0.530 for Cr, 0.219 and 0.003~0.853 for Pb, and 0.088 and 1.73${\times}$10$^{-5}$~0.303 for Fe. These data indicate that the difference between HPE and HPEM is large in the order of Fe, Cr, Pb, Cu, Mn, Cd and Zn. The amounts of heavy metals extracted decreases in the follow order; Sum III(sum of fraction I, II, III in sequential extraction)>HPEM>Sum III (sum of fraction I and II)>HPE for Zn, Cd and Mn and Sum III>HPEM>HPE for Cr and Fe. In the case Cr, Sum II is lower than HPEM and higher than HPE. In case of Cu, extracted heavy metals is large in the order Sum IV>HPEM>Sum III HPE. HPE/HPEM value decreases with increasing the amount of HCl used for maintaining 0.1 N of extraction solution. For samples with high buffer capacity, HPE/HPEM value in all elements is lower than 0.2. On the other hand, for samples with low buffer capacity, HPE/HPEM value are over 0.2 and many samples have values higher than 0.6 for Zn, Cd Mn and Cu due to the small difference between Sum II and Sum III, and relatively higher mobility. However, for Fe and Cr, HPE/HPEM value is below 0.2 even for samples with low buffer capacity due to their low mobility and big difference between Sum II and Sum III. This study indicates that the partial extraction method in Korean Standard Method of soil is not suitable for an assessment of soil contamination in area where buffer capacity of soil can be decreased or lost because of a long term exposure to environmental damage such as acidic rain.

• Asian-Australasian Journal of Animal Sciences
• /
• v.18 no.2
• /
• pp.259-266
• /
• 2005

Steers (n=335) of known genetic backgrounds from four fundamentally different growth types were subjected to two production systems to study differences in carcass traits. Growth types were animals with genetic potential for large mature weight-late maturing, intermediate mature weight-late maturing, intermediate mature weight-early maturing and small mature weight-early maturing. Each year, in a nine-year study, calves of each growth type were weaned and five steers of each growth type were developed on pasture or feedlot and slaughtered at approximately 20 and 14 months of age, respectively. Data collected were pre-slaughter shrunk body weight (SBW); hot carcass weight (HCW); dressing percentage (DRESS); fat thickness at the $12^{th}$ and $13^{th}$ rib interface (FAT); percentage kidney, pelvic, and heart fat (KPH); longissimus muscle area (LMA); marbling score (MARB); quality grade (QG); and yield grade (YG). Year and growth type were significant for all carcass traits. The growth type${\times}$production system interaction was an important source of variation in SBW, HCW; FAT, YG and MARB. The same interaction was non-significant for DRESS, KPH, LMA and QG. Carcass differences in measures of fatness were greater in the feedlot system than in the pasture system. These data could aid producers in matching beef growth type to the production system most suitable for efficient use of resources.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.36 no.2
• /
• pp.143-150
• /
• 2023

Underwater wireless communication is a challenging issue for realizing the smart aqua-farm and various marine activities for exploring the ocean and environmental monitoring. In comparison to acoustic and radio frequency technologies, the visible light communication is the most promising method to transmit data with a higher speed in complex underwater environments. To send data at a speedier rate, high-performance photodetectors are essentially required to receive blue and/or cyan-blue light that are transmitted from the light sources in a light-fidelity (Li-Fi) system. Here, we fabricated high-performance organic phototransistors (OPTs) based on P-type donor polymer (PTO2) and N-type acceptor small molecule (IT-4F) blend semiconductors. Bulk-heterojunction (BHJ) PTO2:IT-4F photo-active layer has a broad absorption spectrum in the range of 450~550 nm wavelength. Solution-processed OPTs showed a high photo-responsivity >1,000 mA/W, a large photo-sensitivity >10³, a fast response time, and reproducible light-On/Off switching characteristics even under a weak incident light. BHJ organic semiconductors absorbed photons and generated excitons, and efficiently dissociated to electron and hole carriers at the donor-acceptor interface. Printed and flexible OPTs can be widely used as Li-Fi receivers and image sensors for underwater communication and underwater internet of things (UIoTs).