• Journal of Preventive Medicine and Public Health
• /
• v.37 no.1
• /
• pp.88-98
• /
• 2004

Objectives : The 3rd community health plan let health centers select and promote core projects considering budget and manpower. This study analyzed the content and selection processes of core projects, using the nationwide 3rd community health plans, to give relevant information on health center policies. Methods : Classification criteria for content analysis of core projects were established and verified through a literature review and by specialist discussions. Fifty plans were selected by stratified proportional random sampling for regional characteristics. And coding criteria standardized through coding repetition and discussion, by 2 persons (k>0.7). Using stratified proportional random sampling for 16 cities and provinces, regional characteristics, 117 plans were selected, and the contents of the core project selection processes and program contents analyzed. Results : The survey was used by 59.8 % of samples as a core project decision-making method. The partici- pants included 98.6, 81.4, 40 and 38.6% of the health staffs, residents, medical institutions, and administrators, respectively. Discussion was used by 15.4% of samples. The participants were health staffs by 100% as a great. The ranking of the frequencies of the selected core projects were, in order; chronic disease control, health promotion, elderly health, maternal-child health, and oral health at 16.4, 14.8, 14.3, 12.7 and 11.9%, respectively. Analyses on the chronic disease control and elderly health contents showed the diversity of object disease, high rates of visitors on patient detection programs, high rates of unclear target populations, and the provision of medical exams and treatments as the main services, with high variations in business per-formance. The national health budgets for health centers in 2003 were about 910 and 240 million won for chronic disease control and elderly health, respectively, which were less than for the other five priority core projects. Conclusions : The chronic disease control and elderly health at the health centers were not standardized for object disease, patient detection program, target population, service provision, and national support budget was insufficient. Thus it is necessary to develop standard guidelines, and increase financial support, for chronic disease control and elderly health

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.4
• /
• pp.161-166
• /
• 2018

In this paper, we propose a method for the random selection of pooling operations for the regularization and reduction of cross validation in convolutional neural networks. The pooling operation in convolutional neural networks is used to reduce the size of the feature map and for its shift invariant properties. In the existing pooling method, one pooling operation is applied in each pooling layer. Because this method fixes the convolution network, the network suffers from overfitting, which means that it excessively fits the models to the training samples. In addition, to find the best combination of pooling operations to maximize the performance, cross validation must be performed. To solve these problems, we introduce the probability concept into the pooling layers. The proposed method does not select one pooling operation in each pooling layer. Instead, we randomly select one pooling operation among multiple pooling operations in each pooling region during training, and for testing purposes, we use probabilistic weighting to produce the expected output. The proposed method can be seen as a technique in which many networks are approximately averaged using a different pooling operation in each pooling region. Therefore, this method avoids the overfitting problem, as well as reducing the amount of cross validation. The experimental results show that the proposed method can achieve better generalization performance and reduce the need for cross validation.

• Asian-Australasian Journal of Animal Sciences
• /
• v.29 no.11
• /
• pp.1530-1540
• /
• 2016

This study was conducted to estimate the genetic parameters of ${\beta}$-hydroxybutyrate (BHBA) and acetone concentration in milk by Fourier transform infrared spectroscopy along with test-day milk production traits including fat %, protein % and milk yield based on monthly samples of milk obtained as part of a routine milk recording program in Korea. Additionally, the feasibility of using such data in the official dairy cattle breeding system for selection of cows with low susceptibility of ketosis was evaluated. A total of 57,190 monthly test-day records for parities 1, 2, and 3 of 7,895 cows with pedigree information were collected from April 2012 to August 2014 from herds enrolled in the Korea Animal Improvement Association. Multi-trait random regression models were separately applied to estimate genetic parameters of test-day records for each parity. The model included fixed herd test-day effects, calving age and season effects, and random regressions for additive genetic and permanent environmental effects. Abundance of variation of acetone may provide a more sensitive indication of ketosis than many zero observations in concentration of milk BHBA. Heritabilities of milk BHBA levels ranged from 0.04 to 0.17 with a mean of 0.09 for the interval between 4 and 305 days in milk during three lactations. The average heritabilities for milk acetone concentration were 0.29, 0.29, and 0.22 for parities 1, 2, and 3, respectively. There was no clear genetic association of the concentration of two ketone bodies with three test-day milk production traits, even if some correlations among breeding values of the test-day records in this study were observed. These results suggest that genetic selection for low susceptibility of ketosis in early lactation is possible. Further, it is desirable for the breeding scheme of dairy cattle to include the records of milk acetone rather than the records of milk BHBA.

• Journal of Korean Society of Forest Science
• /
• v.111 no.2
• /
• pp.302-310
• /
• 2022

Here we aimed to classify the major coniferous tree species (Pinus densiflora, Pinus koraiensis, and Larix kaempferi) by tree measurement information and machine learning algorithms to establish an efficient forest management plan. We used national forest monitoring information amassed over nine years for the measurement information of trees, and random forest (RF), XGBoost (XGB), and light GBM (LGBM) as machine learning algorithms. We compared and evaluated the accuracy of the algorithm through performance evaluation using the accuracy, precision, recall, and F1 score of the algorithm. The RF algorithm had the highest performance evaluation score for all tree species, and highest scores for Pinus densiflora, with an accuracy of about 65%, a precision of about 72%, a recall of about 60%, and an F1 score of about 66%. The classification accuracy for the dominant trees was higher than about 80% in the crown classes, but that of the co-dominant trees, the intermediate trees, and the overtopper trees was evaluated as low. We consider that the results of this study can be used as reference data for decision-making in the selection of thinning trees for forest management.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.56 no.4
• /
• pp.411-418
• /
• 2023

Genome wide association studies (GWAS) identify genetic loci associated with quantitative traits in genomic selection. Although several studies have compared performance of various algorithms, no study compares them in olive flounder Paralichthys olivaceus. This study compared the GWAS results of four mixed linear model (MLM) algorithms and one Fixed and random model Circulating Probability Unification (FarmCPU) algorithm in olive flounder. Considering gender and genetic association matrices as fixed and random effects, the MLM had stable performance without inflation for λ_GC (genomic inflation factor) of -log₁₀P. The FarmCPU algorithm had some appropriate λ_GC of -log₁₀P, and an upward tail was identified in quantile-quantile plots. Therefore, the models were suitable for detecting genetic variants associated with olive flounder growth traits. Moreover, significant genotypes appeared several times at chromosome 22, around which quantitative trait loci are expected to exist. Finally, in both models, some of the most genetic variants were found in genes related to growth traits, confirming their reliability. These results will be helpful when applied to the genomic selection of olive flounder growth traits in the future.

• Journal of Intelligence and Information Systems
• /
• v.29 no.2
• /
• pp.241-265
• /
• 2023

A corporate insolvency prediction model serves as a vital tool for objectively monitoring the financial condition of companies. It enables timely warnings, facilitates responsive actions, and supports the formulation of effective management strategies to mitigate bankruptcy risks and enhance performance. Investors and financial institutions utilize default prediction models to minimize financial losses. As the interest in utilizing artificial intelligence (AI) technology for corporate insolvency prediction grows, extensive research has been conducted in this domain. However, there is an increasing demand for explainable AI models in corporate insolvency prediction, emphasizing interpretability and reliability. The SHAP (SHapley Additive exPlanations) technique has gained significant popularity and has demonstrated strong performance in various applications. Nonetheless, it has limitations such as computational cost, processing time, and scalability concerns based on the number of variables. This study introduces a novel approach to variable selection that reduces the number of variables by averaging SHAP values from bootstrapped data subsets instead of using the entire dataset. This technique aims to improve computational efficiency while maintaining excellent predictive performance. To obtain classification results, we aim to train random forest, XGBoost, and C5.0 models using carefully selected variables with high interpretability. The classification accuracy of the ensemble model, generated through soft voting as the goal of high-performance model design, is compared with the individual models. The study leverages data from 1,698 Korean light industrial companies and employs bootstrapping to create distinct data groups. Logistic Regression is employed to calculate SHAP values for each data group, and their averages are computed to derive the final SHAP values. The proposed model enhances interpretability and aims to achieve superior predictive performance.

• Korean Journal of Logic
• /
• v.19 no.3
• /
• pp.369-400
• /
• 2016

If we learn that 'Mrs Lee has two children and at least one of them is a daughter', what is our credence that her two children are all girls? Obviously it is 1/3. By assuming some other obvious theses it seem to be argued that our credence is 1/2. Also by just supposing we learn trivial information about the future, it seem to be argued that we must change our credence 1/3 into 1/2. However all of these arguments are fallacious, cannot be sound. When using the conditionalization rule to evaluate conformation of a hypothesis by an evidence, or to estimate credence change by information intake, there are some points to keep in mind. We must examine whether relevant information was given through a random procedure or a biased procedure. If someone with full information releases to us particular partial information, an observation, a testimony, an evidence selected intentionally by him, which means the particular partial information was not given by chance, or was not given accidentally or naturally to us, then the conditionalization rule should be employed very cautiously or restrictedly.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.43 no.3 s.309
• /
• pp.31-40
• /
• 2006

Many researches on path planning and obstacle avoidance for the fundamentals of mobile robot have been done. Informations from various sensors can find obstacles and make path. In spite of many solutions of finding optimal path, each can be applied to only a constrained condition. This means that it is difficult to find a universal algorithm. A optimal path with a complicated computation generates a time delay which cannot avoid moving obstacles. In this paper, we propose the algorithm of path planning and obstacle avoidance for mobile robot. We call the proposed method Random Access Sequence(RAS) method. In the proposed method, a small region is set first and numbers are assigned to its neighbors, then the path is selected using these numbers. It has an advantage of fast planning and simple operation. This means that new path selection may be possible within short time and that helps a robot to avoid obstacle in any direction. When a robot meets moving obstacles, it avoids obstacles in a random direction. RAS method using obstacle information from variable sensors is useful to get minimum path length to goal.

• Composites Research
• /
• v.12 no.4
• /
• pp.42-54
• /
• 1999

FMLC(Fiber-Metal Laminate Composites) is a new structural material combining thin metal laminate with adhesive fiber prepreg, it nearly include all the advantage of metallic materials, for example: good plasticity, impact resistance, processibility, light weight and excellent fatigue properties. This research studied the optimum design of the FMLC subject to various loading conditions using genetic algorithm. The finite element method based on the shear deformation theory was used for the analysis of FMLC. Tasi-Hill failure criterion and Miser yield criterion were taken as fitness functions of the fiber prepreg and the metal laminate, respectively. The design variables were fiber orientation angles. In genetic algorithm, the tournament selection and the uniform crossover method were used. The elitist model was also used to be effective evolution strategy and the creeping random search method was adopted in order to approach a solution with high accuracy. Optimization results were given for various loading conditions and compared with CFRP(Carbon Fiber Reinforced Plastic). The results show that the FMLC is more excellent than the CFRP in point and uniform loading conditions and it is more stable to unexpected loading because the deviation of failure index is smaller than that of CFRP.

• Journal of KIISE:Software and Applications
• /
• v.37 no.11
• /
• pp.837-844
• /
• 2010

Prediction of the prosodic phrase boundary is one of the most important natural language processing tasks. We propose, for the natural prediction of the Korean prosodic phrase boundary, a statistical approach incorporating efficient learning features. These new features reflect the factors that affect generation of the prosodic phrase boundary better than existing learning features. Notably, moreover, such learning features, extracted according to the hand-crafted prosodic phrase boundary prediction rule, impart higher accuracy. We developed a statistical model for Korean prosodic phrase boundaries based on the proposed new features. The results were 86.63% accuracy for three levels (major break, minor break, no break) and 81.14% accuracy for six levels (major break with falling tone/rising tone, minor break with falling tone/rising tone/middle tone, no break).