• 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 Remote Sensing
• /
• v.38 no.5_3
• /
• pp.939-951
• /
• 2022

The most important factor in designing autonomous driving systems is to recognize the exact location of the vehicle within the surrounding environment. To date, various sensors and navigation systems have been used for autonomous driving systems; however, all have limitations. Therefore, the need for high-definition (HD) maps that provide high-precision infrastructure information for safe and convenient autonomous driving is increasing. HD maps are drawn using three-dimensional point cloud data acquired through a mobile mapping system (MMS). However, this process requires manual work due to the large numbers of points and drawing layers, increasing the cost and effort associated with HD mapping. The objective of this study was to improve the efficiency of HD mapping by segmenting semantic information in an MMS point cloud into six classes: roads, curbs, sidewalks, medians, lanes, and other elements. Segmentation was performed using various machine learning techniques including random forest (RF), support vector machine (SVM), k-nearest neighbor (KNN), and gradient-boosting machine (GBM), and 11 variables including geometry, color, intensity, and other road design features. MMS point cloud data for a 130-m section of a five-lane road near Minam Station in Busan, were used to evaluate the segmentation models; the average F1 scores of the models were 95.43% for RF, 92.1% for SVM, 91.05% for GBM, and 82.63% for KNN. The RF model showed the best segmentation performance, with F1 scores of 99.3%, 95.5%, 94.5%, 93.5%, and 90.1% for roads, sidewalks, curbs, medians, and lanes, respectively. The variable importance results of the RF model showed high mean decrease accuracy and mean decrease gini for XY dist. and Z dist. variables related to road design, respectively. Thus, variables related to road design contributed significantly to the segmentation of semantic information. The results of this study demonstrate the applicability of segmentation of MMS point cloud data based on machine learning, and will help to reduce the cost and effort associated with HD mapping.

• Journal of Environmental Science International
• /
• v.17 no.5
• /
• pp.573-580
• /
• 2008

This study aims to measure and to analyze the characteristics of thermal environment of the various permeable pavement materials such as a break stone pavement (Green block cubic), soil protection pavement (Soil tector), soil cement pavement and ceramic brick pavement under the summer outdoor environment. The thermal environment characteristics measured in the study includes the changes of surface temperature during the day, and long and short wave radiation of each pavement surface. The experimental condition is based on the data on the hottest temperature (August 9, 2006, $37.1^{\circ}C$) of the year. The albedo was the highest on the break stone pavement(0.8) from 12:00 to 14:00. The albedo of the ceramic brick pavement, a soil tector pavement and soil cement pavement were 0.35, 0.29 and 0.27 from 12:00 to 14:00, respectively. The peak surface temperature and long wave radiation was the highest on the soil protection pavements($56.6^{\circ}C$/627 W/$m^2$). The peak surface temperatures and long wave radiation on the ceramic brick pavement, a stone brick pavement and soil cement pavement were $51.7^{\circ}C$/627 W/$m^2$, $48.8^{\circ}C$/607 W/$m^2$ and $45.9^{\circ}C$/582 W/$m^2$, respectively. The heat environment was better on the break stone pavement than on the other pavements. This is mainly due to the high albedo of the break stone pavement(0.8) while the albedo value of a ceramic brick pavement, a soil tactor pavement and soil cement pavement were 0.35. 0.29 and 0.27. Large heat capacity($2,629kJ/m^3{\cdot}K$) of the stone brick pavements also contributes to this difference. The heat environment was better on the soil cement pavement than the soil tector pavement. This is mainly due to the evaporation of the soil cement pavement while the active evaporation of the soil tactor pavement was not continued after two days from the rainfall event. To improve the thermal environments in the urban area, it is recommended to raise the albedo of the pavements by brightening the surface color of the pavement materials. Further studies on the pavement materials and the construction methods which can enhance the continuous evapotranspiration from the pavements surface are needed.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.38 no.3
• /
• pp.115-125
• /
• 2010

A theme park is not just a recreational space for leisure activities, but also a place of storytelling as collected around abstract boundaries called themes. These stories are 'a space that tells the meaning' that the visitor is looking for and the Robot Land space offers robots, humans, and nature. This study is a description of the design strategy and content of the work which was elected as a subject of the subsequent rank negotiation of the Masan Robot Land design contest for the selection of a private contractor. The focus of the plan is, first, the organizational power of each space and the delivery power of a theme for the history of revisits, which might be considered depending on whether or not the theme park has been successful in the visitor's mind. Second, it is to actively use the potential of Masan, which is not only the key hub of the mechanical industry but also has beautiful coastal resources. First, they created a space that can flexibly react depending on the user's desire and the change of form, minimizing environmental damage by using a linear metabolism that can provide an amalgam of the elemental characteristics of robots, humans, and nature as motifs. They introduced a planting plan for the admissions square, an existing forest, slope, vacation spot, the inside of a complex, and Eco Island, etc. by utilizing symbolic meaning and adjusting to the spatial characteristics of each space. In addition, they sought a detailed space by setting up zones tailored to the use and character of the subject area, having exhibitions and education about robots, vacation facilities for lodgers, various recreational and commercial facilities, and space for utopian gardens as themes. They planned Masan Robot Land to be a true cultural space that creates mental richness on the basis of not only the economical effects but also local emotion.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.62 no.6
• /
• pp.85-95
• /
• 2020

According to the standard guidelines of design flood (MLTM, 2012; MOE, 2019), the design flood is calculated based on past precipitation. However, due to climate change, the frequency of extreme rainfall events is increasing. Therefore, it is necessary to analyze future floods' volume by using climate change scenarios. Meanwhile, the standard guideline was revised by MOE (Ministry of Environment) recently. MOE proposed modified Huff distribution and new CN (Curve Number) value of forest and paddy. The objective of this study was to analyze the change of flood volume by applying the modified Huff and newly proposed CN to the probabilistic precipitation based on SSP and RCP scenarios. The probabilistic rainfall under climate change was calculated through RCP 4.5/8.5 scenarios and SSP 245/585 scenarios. HEC-HMS (Hydrologic Engineering Center - Hydrologic Modeling System) was simulated for evaluating the flood volume. When RCP 4.5/8.5 scenario was changed to SSP 245/585 scenario, the average flood volume increased by 627 ㎥/s (15%) and 523 ㎥/s (13%), respectively. By the modified Huff distribution, the flood volume increased by 139 ㎥/s (3.76%) on a 200-yr frequency and 171 ㎥/s (4.05%) on a 500-yr frequency. The newly proposed CN made the future flood value increase by 9.5 ㎥/s (0.30%) on a 200-yr frequency and 8.5 ㎥/s (0.25%) on a 500-yr frequency. The selection of climate change scenario was the biggest factor that made the flood volume to transform. Also, the impact of change in Huff was larger than that of CN about 13-16 times.

• Journal of the Korean Institute of Traditional Landscape Architecture
• /
• v.31 no.1
• /
• pp.13-27
• /
• 2013

This study investigated flora and vascular plants area around Buyongdae where include optimum view point Okyeonjeongsa and Gyeomamjeongsa in Hahoe Village, Andong. Based on the findings, the aim of results was attempted to suggest the management plan of the vegetation area around Buyongdae which designated UNESCO World cultural heritage and folk village in Korea. The results of the study is as the following. All flora of this site, Buyongdae in Andong, were 301 taxa; 89 families, 217 genera, 251 species, 1 subspecies, 38 varieties and 11 forms. Vegetation of each of areas in site was classified Pinus densiflora as representative forest around Buyongdae, Quercus variabilis forest around Gyeomamjjeongsa and Koelreuteria paniculata forest in river cliff of Buyongdae. The 14 taxa(Polygala tenuifolia, Koelreuteria paniculata, and others) as the rare plants and the 7 taxa(Weigela subsessilis, and others) as endemic species were recorded in the surveyed site among the whole flora. The specific plants which is categorized to degree by the Ministry of Environment appeared as 32 taxa; degree IV was 4 species; Polygala tenuifolia, etc. and degree III was 10 species; Hypodematium glandulosopilosum, Pyrrosia petiolosa, etc. and degree I was 13 species; Hemiptelea davidii, Zizyphus jujuba var. jujuba, etc. Also, the 11 taxa were recorded in limestone area as the chamaephyte; Hypodematium glandulosopilosum, Celtis koraiensis, Siphonostegia chinensis, Artemisia gmelini, and others. The naturalized plants recorded as 25 taxa; Viola arvensis, Erechtites hieracifolia, etc. and invasion of Sicyos angulatus which is ecosystem disturbing plants. The naturalization rate(NR) was 8.3% and the urbanization index(UI) was 7.8% by field survey. For recovery of indigenous to traditional vegetation landscape in Buyongdae, Pinus rigida of upper-growth and Alnus sibirica of middle-growth need thinning. In addition, the site demands both sustainable management through long-term monitering and gradual elimination method for the naturalized plants including Sicyos angulatus where found in Kyumamjeongsa. Particularly, intentional planting for decoration and cultivation around Hwachunseowon, Helianthus tuberosus must be substitute with the native plants. Meanwhile, Polygala tenuifolia is potently demanded preventing habitate, confirming additional habitate, seed securing and preservation of gene resource internally or externally. Koelreuteria paniculata; denoting feature of river cliff, including Hypodematium glandulosopilosum, Siphonostegia chinensis, Zizyphus jujuba var. jujuba, Pyrrosia petiolosa, and Celtis koraiensis, also need an assertive preservation. Futhermore, Wisteria floribunda for. floribunda, located around Galmo rock to Chunggil course, remained by colonizing form, need preservation and observation for a while. Beside that, the information boards are required to educate visiting user about safe regulations on the narrow way in rock hill.

• Journal of Korean Society of Forest Science
• /
• v.42 no.1
• /
• pp.74-82
• /
• 1979

Pinus densiflora S. et Z. which has the widest distribution and highest stock at present in Korea and has used as materials for fuel, house construction, furniture and many other purposes for several hundred years, is considered as one of major species of economic importance, although there are a few epidemic insect injuries and silvicultural difficulties in regeneration. However, since disorderly cutting has been conducted for long time, the valuable local stands have disappeared. Therefore immediate attempts should be taken on the gene conservation and genetic studies including heritabilities and genetic gains of desirable characters. One hundred and twently five plus trees have so far been selected from the nationwide area and kept for the purpose of seed orchard establishment and other theoretical studies. In this study, the wind pollinated seeds of grafted stocks of 13 plus trees in clone bank located in Suweon and the seeds of 4 Japanese plus trees were collected and their progenies were used for heritability study. As indicated in figure 1, thirteen plus trees are from the middle part of Korea and two experimental plantations were laid out in ]972 (1-1 stock) by the randomized block design with five replications, consisting of 10 trees line plots of each family in each replication. Tree height, root collar diameter and branch diameter at 1cm apart from the main stem were measured. The results are summarized as follows 1. The rank of height growth of each plus tree progenies by age was greatly changed under age 3 and it was affected more in poor site than good site. 2. The heritabilities of height growth were estimated to be 7.2% at age 3, negative sign at age 4, 9.4% at age 5, 13.0% at age 6, 8.1% at age 7 and abrupt increase of 63.8% at age 8. The heritabilities were generally increased with increase of age. 3. The heritabilities of root collar diameter and branch diameter (average of the three biggest branches) was 3.2% and 11.8% at age 8 respectively. 4. The genetic gain was largest at age 8 and it was 46.6% when a selection intensity, 1/500 was taken. It, therefore, seems to be reasonable that selection should not be made under age 7.

• Journal of Korean Society of Forest Science
• /
• v.84 no.2
• /
• pp.239-246
• /
• 1995

This research aimed at the contribution to obtaining the scientifical data which were required for planning she environmentally sound and sustainable management, particularly in the field of the logging road construction. Main natural environmental variables including natural vegetation, rainfall, soil runoff were measured in the logging road on-sites and analysed. This project was carried out at the (mt.)Paekunsan Research sorest of Seoul National University, located in Gwangyang, Chollanam-do in southern part of Korea, from 1993 to 1994. 1. The explanatory variables for erosion and sedimentation on logging road surface were accumulated rainfall, erosion distance, cross-sectional gradient, and soil hardness. The erosion and sedimentation on logging road was increasing positively in proportion to the accumulated rainfall, soil distance from starting point of the logging road, and cross-sectional gradient. 2. On cut-slope of logging road, cut-slope shape, part of the slope, plant coverage, soil hardness, sand content, accumulated rainfall, clay content, and silt content were effective factors. Cut-slope erosion and sedimentation on logging roam increased as with the lower plant coverage, the lower accumulated rainfall, the high sand content in the soil. 3. On fill-slope of logging road, there were three significant variables such as total rainfall and number of rainfall-storm. Fill-slope erosion and sedimentation had a positive correlation with the amount of rainfall, the number of rainfall, the soil hardness. 4. The total erosion and sedimentation on logging road were $5.04{\times}10^{-2}m^2/m^2$ in logging road construction year, $7.37{\times}10^{-2}m^2/m^2$ in next year. The erosion and sedimentation on logging road surface were 32.7% of total erosion and sedimentation on Logging road in construction year, and 57.1% in next year, respectively. The erosion and sedimentation on cut-slopes were 30.4% on logging road in construction year, fill-slopes of total erosion and sedimentation and 21.0% in next year, respectively. The erosion and sedimentation on fill-slopes were 36.9% on logging road in construction year, 21.9 in next year. To decrease the erosion and sedimentation at the logging road from the beginning stage of construction, the effective revegetation works should be implemented on the cut-slope and fill slopes, and erosion control measures such as optima. road design must be constructed on read surface.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.9 no.1
• /
• pp.7-14
• /
• 1989

This experiment was carried out to determine the seasonal herbage production and utilization during the growing season of pasture on the forest (shading 30%). Plant height, leaf area index(LAl), dry matter(DM) production and distribution, chemical composition, in vitro dry matter digestibility(IVDMD), herbage utilization percentage and chewing efficiency were investigated using the Corridale sheep. Experimental field was treated by one plot design(3 rep.) and performed from 1987 to 1988 at Chungnam National University, Daejon. The results obtained are summarized as follows: 1. The highest plant height and LA1 were observed in May(35.0 cm, 4.89), followed by April(28.0 cm, 4.23), while the plant height and LA1 in October (13.0 cm, 0.49) showed very low. 2. During the growing season, about 58.3 % of annual DM production (7240 kg/ha) was produced during the spring (April, May and June) and the highest DM production was obtained in May (2040 kg/ha), which was more than 28.2 % of total DM production. However, DM production in July and August was about 24.2 % and those in September and October (17.5 %) was very low, but the difference of DM production from June to September was small. 3. The maximum DM production per day (65.8 kg/ha) was observed in May, followed by June (28.7 kglha), while DM production per day in October (16.5 kg/ha) showed very low (p <0.01). 4. Crude protein content and IVDMD of herbage samples during the spring (April, May and June) were higher, while crude fiber, ADF, and NDF content were lower in an summer growth herbage samples (July and August), but autumn growth herbage samples was intermediate. Crude ash content and IVDMD of collected herbage samples were slightly more, while crude fiber, ADF and NDF content were slightly less than offered and residued herbage samples during the growing season. 5. The maximum DM intake per metabolic body size was observed in May(68.9 g), followed by October (66.7 g), while very low in August (52.5 g). Significant positive correlation (p <0.05) was found between DM intake and IVDMD. 6. Herbage utilization percentage was very high in April (83.4 %), while very low in August (64.0 %). The percentage of annual herbage utilization was about 75.5 %. 7. The maximum ruminating and chewing efficiency of herbage samples were observed in May, followed by October, while very low in August.

• Journal of Korean Society of Forest Science
• /
• v.77 no.3
• /
• pp.259-268
• /
• 1988

For the purpose of examining the effects of simulated acid rain on nutrient contents of plant tissues in Pintos densiflora seedlings and Forsythia koreana rooted cuttings, the experimental design of randomized block arrangement with three replications was implemented in the experimental field of Yesan National Agricultural Junior College. One-year-old Pinus densiflora seedlings and Forsythia koreana cuttings were planted in the pots filled the mixed soils(nursery soil : forest soil of siliceous sandy loam=1 : 1 v/v) in the early spring of 1986. The regime of artificial acid rain, in terms of spray frequency per month and spray amount at single treatment per plot, was simulated on the basis of climatological data averaged from 30 years records. Simulated acid rain(pH 2.0, pH 4.0, and pH 5.5 as control) containing sulfuric and nutric acid in the ratio of 3 : 2(chemical equivalant basis) diluted with ground water were treated on the experimental plants under condition of cutting off the natural precipitation with vinyl tunnel, during the growing season(May 1 to August 31) in 1986. The results obtained in this study were as follow : (1) As for the nitrogen contents in plant tissues, P. densiflora increased significantly in one-year-old stembranch and root tissues, and F. koreana increased significantly in leaf and root tissues, as the pH levels of acid rain decreased. (2) The available phosphate contents in root tissues of P. densiflora, and in leaf and root tissues of F. koreana were significantly decreased, as the pH levels of acid rain decreased. (3) $K_2O$, CaO and MgO contents in plant tissues were significantly decreased in the both species as the pH levels decreased. And the effects of acid rain on F. koreana were higher than those of P. densiflora. (4) Sulfur contents of plant tissues in the both species were increased at pH 2.0 treatment. There were significant differences among three acid rain treatments in leaf and root tissues of P. densiflora, and in all parts of F. koreana. (5) In the effects of simulated acid rain on the both species and the tested soils, in general, F. koreana revealed higher sensitiveness than P. densiflora, and the lower pH levels of simulated acid rain were treated, the more sharp reaction was showed.