• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.6
• /
• pp.431-440
• /
• 2021

The purpose of this study was to classify the latent profiles of elementary school teachers' job stress and to explore the effects of the relative variables to determine these classifications. In addition, the differences in the teacher-student relationship and parental involvement in school based on the classification were discussed. Data from 709 elementary school teachers who participated in the 11th wave of the Panel Study on Korean Children in 2018 were analyzed by Latent Profile Analysis (LPA). The findings can be summarized as follows. First, four subgroups could be defined according to the elementary school teachers' job stress: low-level job stress group, mid-level job stress group, mid-level administrative work stress group, and mid-level relationship and guidance stress group. Second, the final education and average time to work were significant determinants of the latent groups. Third, teacher-student conflict and parental involvement in school showed differences between the subgroups. Specifically, the mid-level relationship and guidance stress group reported the highest conflict level with children and the lowest parental involvement in school. These findings suggest promoting relief and preventative training programs for elementary school teachers to overcome various job stress.

• Journal of The Korean Society of Clinical Toxicology
• /
• v.18 no.2
• /
• pp.78-84
• /
• 2020

Purpose: In acute acetaminophen poisoning, the administration of N-acetylcysteine (NAC) can effectively treat the main complications, such as kidney injury and liver failure. In the current situation, measurements of the acetaminophen concentration are not checked in the usual medical facilities. Therefore, this study examined the factors of determining the administration of NAC in addition to the stated amount of intake. Methods: The medical records of patients who visited Ajou University Hospital emergency center with acetaminophen poisoning from January 2015 to December 2019 were reviewed retrospectively. One hundred and seventy-nine patients were initially included. Among these patients, 82 patients were finally selected according to the inclusion criteria in the study. The inclusion criteria were as follows: patients who were 15 years of age or older; those whose ingested dose, ingested time, and body weight were clearly identified; and patients whose acetaminophen sampling time was within 24 hours. Patients were divided into two groups: NAC administered vs. non-NAC administered. The following variables were compared in these two groups: ingested dose, ingested dose per body weight, hospital arrival time after ingestion, suicide attempt history, psychiatric disease history, classification of toxic/non-toxic groups, duration of hospitalization, and laboratory results. Results: Univariate analysis revealed the ingested dose per body weight, hospital arrival time after ingestion, suicide attempt history, and psychiatric disease history to be the determining factors in administering NAC. Logistic regression analysis confirmed that the ingested dose per body weight was the only significant factor leading to an NAC treatment decision. (Odds ratio=1.039, 95% Confidential interval=1.009-1.070, p=0.009) Conclusion: The ingested dose per body weight was the only determining factor for administering NAC in patients with acute acetaminophen poisoning. On the other hand, additional criteria or indicators for the NAC administration decision will be necessary considering the inaccuracy of the ingested dose per body weight and the efficiency of NAC administration.

• Journal of Intelligence and Information Systems
• /
• v.28 no.2
• /
• pp.307-332
• /
• 2022

One of the most intensively conducted research areas in business application study is a bankruptcy prediction model, a representative classification problem related to loan lending, investment decision making, and profitability to financial institutions. Many research demonstrated outstanding performance for bankruptcy prediction models using artificial intelligence techniques. However, since most machine learning algorithms are "black-box," AI has been identified as a prominent research topic for providing users with an explanation. Although there are many different approaches for explanations, this study focuses on explaining a bankruptcy prediction model using a counterfactual example. Users can obtain desired output from the model by using a counterfactual-based explanation, which provides an alternative case. This study introduces a counterfactual generation technique based on a genetic algorithm (GA) that leverages both domain knowledge (i.e., causal feasibility) and feature importance from a black-box model along with other critical counterfactual variables, including proximity, distribution, and sparsity. The proposed method was evaluated quantitatively and qualitatively to measure the quality and the validity.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_3
• /
• pp.967-977
• /
• 2022

The occurrence and intensity of wildfires are increasing with climate change. Emissions from forest fire smoke are recognized as one of the major causes affecting air quality and the greenhouse effect. The use of satellite product and machine learning is essential for detection of forest fire smoke. Until now, research on forest fire smoke detection has had difficulties due to difficulties in cloud identification and vague standards of boundaries. The purpose of this study is to detect forest fire smoke using Level 1 and Level 2 data of Geostationary Environment Monitoring Spectrometer (GEMS), a Korean environmental satellite sensor, and machine learning. In March 2022, the forest fire in Gangwon-do was selected as a case. Smoke pixel classification modeling was performed by producing wildfire smoke label images and inputting GEMS Level 1 and Level 2 data to the random forest model. In the trained model, the importance of input variables is Aerosol Optical Depth (AOD), 380 nm and 340 nm radiance difference, Ultra-Violet Aerosol Index (UVAI), Visible Aerosol Index (VisAI), Single Scattering Albedo (SSA), formaldehyde (HCHO), nitrogen dioxide (NO₂), 380 nm radiance, and 340 nm radiance were shown in that order. In addition, in the estimation of the forest fire smoke probability (0 ≤ p ≤ 1) for 2,704 pixels, Mean Bias Error (MBE) is -0.002, Mean Absolute Error (MAE) is 0.026, Root Mean Square Error (RMSE) is 0.087, and Correlation Coefficient (CC) showed an accuracy of 0.981.

• The Journal of the Korea Contents Association
• /
• v.21 no.7
• /
• pp.340-350
• /
• 2021

This study was conducted by utilizing the Q research method, which is one of the qualitative analysis methods that can approach the in-depth and essential meaning of consumers' restaurant O2O service operation behavior. The purpose of this study is to classify the behavior of restaurant O2O services by type, to find out the characteristics of variables, and to suggest future improvement directions. An exploratory study was conducted using the Q-methodology to analyze the subjective perception of the restaurant O2O service behavior. To this end, positive and negative statement cards were prepared, P samples were selected, and Q-sort, which was subjected to classification, was analyzed using the PC QUANL program and Q factor analysis. As a result of the analysis, it was classified into three single types. Type 1 【(N= 7: Restaurant O2O Service Convenience Syndrome Type】, Type 2 【(N= 7): Restaurant O2O Service Benefit Pursuit Type】, Type 3 【(N= 6): Restaurant O2O Service Convenience Type】 The name of the factor was set as [Type], and it was found that each type has different characteristics. Through this analysis, the marketing strategy according to each factor detected is presented, and the point of supplementing the restaurant's O2O service and the direction of future operation. services in future studies.

• Journal of People, Plants, and Environment
• /
• v.23 no.3
• /
• pp.267-276
• /
• 2020

Background and objective: Agro-healing is crucial with urban farming in the domestic. It is beyond the conventional agriculture. This study was carried out to assess the 2010s researches on domestic agro-healing and predict the future direction of agro-healing development. Methods: Among the articles published from 2010 to 2018, we collected some articles by searching keywords including agro-healing, garden activity, gardening, horticultural activity, horticultural program, horticultural therapy, plant effect, plant environment, plant growing program, plant impact, social gardening, urban agriculture and vegetable garden activity, selected 83 articles that were evaluated in advance, and analyzed by frequency analysis, t-test, and one-way ANOVA with SPSS 20.0. Results: Agro-healing journal articles were published the most in 2010, and have declined since then. In the classification according to the academic society, most of the journal articles were published by the Society for People, Plants, and Environment. The main targets of domestic agro-healing related to activities and programs were preschoolers, children and adolescents, accounting for 52.4% of the total. By the characteristics of the subjects, agro-healing programs and studies were conducted with special participants who needed special care compared to the general participants. The dependent variables were classified into six areas according to their attributes and the share of psychological and emotional areas was highest (42.6%) among them. In terms of the composition of the program, the share of those with 9-12 sessions was highest (36.7%) and the share of those with more than 20 participants was also highest (39.8%). Conclusion: It is recommended to operate agro-healing programs or industries focusing on the socially disadvantaged including those who have special needs or the underprivileged, but in order to create income for farms and expand the demand for agro-healing, it will be necessary to spread the perception that anyone without any physical or emotional issue can be the targets of and experience agro-healing. To meet the different needs of targets of agro-healing, it will be necessary to conduct objective and practical studies on broader areas and in the process the healing functions of agriculture and the strength of agro-healing needs to be further highlighted.

• The Journal of the Korea Contents Association
• /
• v.21 no.11
• /
• pp.135-144
• /
• 2021

Recently, data-driven decision-making technology has become a key technology leading the data industry, and machine learning technology for this requires high-quality training datasets. However, real-world data contains missing values for various reasons, which degrades the performance of prediction models learned from the poor training data. Therefore, in order to build a high-performance model from real-world datasets, many studies on automatically imputing missing values in initial training data have been actively conducted. Many of conventional machine learning-based imputation techniques for handling missing data involve very time-consuming and cumbersome work because they are applied only to numeric type of columns or create individual predictive models for each columns. Therefore, this paper proposes a new data imputation technique called 'Denoising Self-Attention Network (DSAN)', which can be applied to mixed-type dataset containing both numerical and categorical columns. DSAN can learn robust feature expression vectors by combining self-attention and denoising techniques, and can automatically interpolate multiple missing variables in parallel through multi-task learning. To verify the validity of the proposed technique, data imputation experiments has been performed after arbitrarily generating missing values for several mixed-type training data. Then we show the validity of the proposed technique by comparing the performance of the binary classification models trained on imputed data together with the errors between the original and imputed values.

• Journal of Korea Entertainment Industry Association
• /
• v.15 no.7
• /
• pp.45-57
• /
• 2021

The purpose of this study is to identify what factors have impacts on the amount of YouTube viewing. In doing so, usage type, children's levels of self-control on viewing, parent's perception of media contents, and parental mediation style were investigated by age. The result of the study showed no significant differences of the amount of use by the different age groups. There were, however, significant differences in terms of subscription status(non subscription based vs. subscription-based viewing), level of self-control, perception of content, and parent's mediations (technology, supervision, and guidance). Given the amount of YouTube use, the subscription status and parent's supervision were significantly influential factors for the age group of 3-4 years old. For the age group of 5-6, subscription status, levels of self-control, and mediation of parent's supervision and guidance were influential factors. For the age of 7-9, subscription status, the level of self control, and premium service were significantly influential. Finding similarities and differences in meaningful variables by age group suggests that different strategies should be used to reduce the amount of children's YouTube use. In addition, it raises the need for a more detailed classification of children's YouTube usage methods, which have not been addressed so far, and the need for research on the influence of these methods.

• Korean Journal of Ecology and Environment
• /
• v.55 no.2
• /
• pp.156-166
• /
• 2022

Understanding the characteristics of reservoir water quality is fundamental in reservoir ecosystem management. The water quality of reservoirs is affected by various factors including hydro-morphology of reservoirs, land use/cover, and human activities in their catchments. In this study, we classified 83 major reservoirs in South Korea based on nine physicochemical factors (pH, dissolved oxygen, chemical oxygen demand, total suspended solid, total nitrogen, total phosphorus, total organic carbon, electric conductivity, and chlorophyll-a) measured for five years (2015~2019). Study reservoirs were classified into five main clusters through hierarchical cluster analysis. Each cluster reflected differences in the water quality of reservoirs as well as hydromorphological variables such as elevation, catchment area, full water level, and full storage. In particular, water quality condition was low at a low elevation with large reservoirs representing cluster I. In the comparison of eutrophication status in major reservoirs in South Korea using the Korean trophic state index, in some reservoirs including cluster IV composed of lagoons, the eutrophication was improved compared to 2004~2008. However, eutrophication status has been more impaired in most agricultural reservoirs in clusters I, III, and V than past. Therefore, more attention is needed to improve the water quality of these reservoirs.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.107-119
• /
• 2023

The technology that recognizes a soldier's motion and movement status has recently attracted large attention as a combination of wearable technology and artificial intelligence, which is expected to upend the paradigm of troop management. The accuracy of state determination should be maintained at a high-end level to make sure of the expected vital functions both in a training situation; an evaluation and solution provision for each individual's motion, and in a combat situation; overall enhancement in managing troops. However, when input data is given as a timer series or sequence, existing feedforward networks would show overt limitations in maximizing classification performance. Since human behavior data (3-axis accelerations and 3-axis angular velocities) handled for military motion recognition requires the process of analyzing its time-dependent characteristics, this study proposes a high-performance data-driven classifier which utilizes the long-short term memory to identify the order dependence of acquired data, learning to classify eight representative military operations (Sitting, Standing, Walking, Running, Ascending, Descending, Low Crawl, and High Crawl). Since the accuracy is highly dependent on a network's learning conditions and variables, manual adjustment may neither be cost-effective nor guarantee optimal results during learning. Therefore, in this study, we optimized hyperparameters using Bayesian optimization for maximized generalization performance. As a result, the final architecture could reduce the error rate by 62.56% compared to the existing network with a similar number of learnable parameters, with the final accuracy of 98.39% for various military operations.