• Korean Journal of Remote Sensing
• /
• v.25 no.4
• /
• pp.311-320
• /
• 2009

This study was carried out to estimate forest biomass and to produce forest biomass thematic map for Muju county by combining field data from the 5$^{th}$ National Forest Inventory (2006-2007) and satellite data. For estimating forest biomass, two methods were examined using a Landsat TM-5(taken on April 28th, 2005) and field data: multi-variant regression modeling and t-Nearest Neighbor (k-NN) technique. Estimates of forest biomass by the two methods were compared by a cross-validation technique. The results showed that the two methods provide comparatively accurate estimation with similar RMSE (63.75$\sim$67.26ton/ha) and mean bias ($\pm$1ton/ha). However, it is concluded that the k-NN method for estimating forest biomass is superior in terms of estimation efficiency to the regression model. The total forest biomass of the study site is estimated 8.4 million ton, or 149 ton/ha by the k-NN technique.

• Asian Spine Journal
• /
• v.12 no.6
• /
• pp.1060-1068
• /
• 2018

Study Design: Retrospective study. Purpose: Lumbar intervertebral disc degeneration is an important cause of low back pain. Overview of Literature: Spinal fusion is often reported to have a good course for adolescent idiopathic scoliosis (AIS). However, many studies have reported that adjacent segment degeneration is accelerated after lumbar spinal fusion. Radiography is a simple method used to evaluate the orientation of the vertebral column. magnetic resonance imaging (MRI) is the method most often used to specifically evaluate intervertebral disc degeneration. The Pfirrmann classification is a well-known method used to evaluate degenerative lumbar disease. After spinal fusion, an increase in stress, excess mobility, increased intra-disc pressure, and posterior displacement of the axis of motion have been observed in the adjacent segments. Methods: we retrospectively secured and analyzed the data of 15 patients (four boys and 11 girls) with AIS who underwent a spinal fusion surgery. We studied the full-length view of the spine (anterior-posterior and lateral) from the X-ray and MRI obtained from all patients before surgery. Postoperatively, another full-length spine X-ray and lumbosacral MRI were obtained from all participants. Then, pelvic tilt, sacral slope, curve correction, and fused and free segments before and after surgery were calculated based on X-ray studies. MRI images were used to estimate the degree to which intervertebral discs were degenerated using Pfirrmann grading system. Pfirrmann grade before and after surgery were compared with Wilcoxon signed rank test. While analyzing the contribution of potential risk factors for the post-spinal fusion Pfirrmann grade of disc degeneration, we used generalized linear models with robust standard error estimates to account for intraclass correlation that may have been present between discs of the same patient. Results: The mean age of the participant was 14 years, and the mean curvature before and after surgery were 67.8 and 23.8, respectively (p<0.05). During the median follow-up of 5 years, the mean degree of the disc degeneration significantly increased in all patients after surgery (p<0.05) with a Pfirrmann grade of 1 and 2.8 in the L2-L3 before and after surgery, respectively. The corresponding figures at L3-L4, L4-L5, and L5-S1 levels were 1.28 and 2.43, 1.07 and 2.35, and 1 and 2.33, respectively. The lower was the number of free discs below the fusion level, the higher was the Pfirrmann grade of degeneration (p<0.001). Conversely, the higher was the number of the discs fused together, the higher was the Pfirrmann grade. Conclusions: we observed that the disc degeneration aggravated after spinal fusion for scoliosis. While the degree of degeneration as measured by Pfirrmann grade was directly correlated by the number of fused segments, it was negatively correlated with the number of discs that remained free below the lowermost level of the fusion.

• Journal of Wetlands Research
• /
• v.23 no.2
• /
• pp.122-132
• /
• 2021

We reviewed the methods employed in Korean tidal flat surveys to measure the local abundance of the endangered wildlife and marine protected species, the fiddler crab, Austruca lactea. A complete census for infinite population is impossible even in a limited habitat within a tidal flat, and density estimates from samples strongly vary due to diverse biological and ecological factors. The habitat boundaries and areas shift with periodicities or rhythmic activities of organisms as well as measurement errors. Hence the local abundance calculated from density and habitat areas should be regarded as transient. This conjecture was valid based on the spatio-temporal variations of the density averages, standard error ranges, and spatial distribution of the crab, A. lactea observed for 3 years (2015-2017) in Songdo tidal flat in Incheon. We proposed the potential habitat areas using the occurrence probability of 50% from logistic regression model, reflecting the importance of habitat conservation value as an alternative to local abundance. The spatial shape of potential habitat predicted from a generalized model would remain constant over time unless the species' critical environmental conditions change rapidly. The species-specific model is expected to be used for the introduction of desired species in future habitat restoration/creation projects.

• Journal of IKEEE
• /
• v.25 no.1
• /
• pp.148-155
• /
• 2021

As robot technology advances, research on the driving system of mobile robots is actively being conducted. The driving system of a mobile robot configured based on two-wheels and four-wheels has an advantage in unidirectional driving such as a straight line, but has disadvantages in turning direction and rotating in place. A ball robot using a ball as a wheel has an advantage in omnidirectional movement, but due to its structurally unstable characteristics, balancing control to maintain attitude and driving control for movement are required. By estimating the position from an encoder attached to the motor, conventional ball robots have a limitation, which causes the accumulation of errors during driving control. In this study, a driving control system was proposed that estimates the position coordinates of a ball robot through image processing and uses it for driving control. A driving control system including an image processing unit, a communication unit, a display unit, and a control unit for estimating the position of the ball robot was designed and manufactured. Through the driving control experiment applying the driving control system of the ball robot, it was confirmed that the ball robot was controlled within the error range of ±50.3mm in the x-axis direction and ±53.9mm in the y-axis direction without accumulating errors.

• Korean Journal of Remote Sensing
• /
• v.39 no.1
• /
• pp.47-63
• /
• 2023

Unmanned aerial vehicle (UAV) and sensor technologies are rapidly developing and being usefully utilized for spatial information-based agricultural management and smart agriculture. Until now, there have been many difficulties in obtaining production information in a timely manner for large-scale agriculture on reclaimed land. However, smart agriculture that utilizes sensors, information technology, and UAV technology and can efficiently manage a large amount of farmland with a small number of people is expected to become more common in the near future. In this study, we evaluated the productivity of forage maize grown on reclaimed land using UAV and sensor-based technologies. This study compared the plant height, vegetation cover ratio, fresh biomass, and dry biomass of maize grown on general farmland and reclaimed land in South Korea. A biomass model was constructed based on plant height, cover ratio, and volume-based biomass using UAV-based images and Farm-Map, and related estimates were obtained. The fresh biomass was estimated with a very precise model (R² =0.97, root mean square error [RMSE]=3.18 t/ha, normalized RMSE [nRMSE]=8.08%). The estimated dry biomass had a coefficient of determination of 0.86, an RMSE of 1.51 t/ha, and an nRMSE of 12.61%. The average plant height distribution for each field lot was about 0.91 m for reclaimed land and about 1.89 m for general farmland, which was analyzed to be a difference of about 48%. The average proportion of the maize fraction in each field lot was approximately 65% in reclaimed land and 94% in general farmland, showing a difference of about 29%. The average fresh biomass of each reclaimed land field lot was 10 t/ha, which was about 36% lower than that of general farmland (28.1 t/ha). The average dry biomass in each field lot was about 4.22 t/ha in reclaimed land and about 8 t/ha in general farmland, with the reclaimed land having approximately 53% of the dry biomass of the general farmland. Based on these results, UAV and sensor-based images confirmed that it is possible to accurately analyze agricultural information and crop growth conditions in a large area. It is expected that the technology and methods used in this study will be useful for implementing field-smart agriculture in large reclaimed areas.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.6
• /
• pp.259-266
• /
• 2023

It is very difficult to measure the performance of the machine learning model in the business service stage. Therefore, managing the performance of the model through the operational department is not done effectively. Academically, various studies have been conducted on the concept drift detection method to determine whether the model status is appropriate. The operational department wants to know quantitatively the performance of the operating model, but concept drift can only detect the state of the model in relation to the data, it cannot estimate the quantitative performance of the model. In this study, we propose a performance prediction model (PPM) that quantitatively estimates precision through the statistics of concept drift. The proposed model induces artificial drift in the sampling data extracted from the training data, measures the precision of the sampling data, creates a dataset of drift and precision, and learns it. Then, the difference between the actual precision and the predicted precision is compared through the test data to correct the error of the performance prediction model. The proposed PPM was applied to two models, a loan underwriting model and a credit card fraud detection model that can be used in real business. It was confirmed that the precision was effectively predicted.

• Journal of the Korea Society for Simulation
• /
• v.18 no.4
• /
• pp.137-148
• /
• 2009

The purpose of this study is to implement variable selection algorithm which helps construct a reliable linear regression model. If we use all candidate variables to construct a linear regression model, the significance of the model will be decreased and it will cause 'Curse of Dimensionality'. And if the number of data is less than the number of variables (dimension), we cannot construct the regression model. Due to these problems, we consider the variable selection problem as a combinatorial optimization problem, and apply GA (Genetic Algorithm) to the problem. Typical measures of estimating statistical significance are $R^2$, F-value of regression model, t-value of regression coefficients, and standard error of estimates. We design GA to solve multi-objective functions, because statistical significance of model is not to be estimated by a single measure. We perform experiments using simulation data, designed to consider various kinds of situations. As a result, it shows better performance than LARS (Least Angle Regression) which is an algorithm to solve variable selection problems. We modify algorithm to solve portfolio selection problem which construct portfolio by selecting stocks. We conclude that the algorithm is able to solve real problems.

• Korean Journal of Construction Engineering and Management
• /
• v.24 no.5
• /
• pp.22-34
• /
• 2023

Predicting accurately the construction cost budget in the early stages of construction projects is crucial to support the client's decision-making and achieve the objectives of the construction project. This holds true for public construction projects as well. However, the current methods for predicting construction cost budgets in the early stages of public construction projects are not sophisticated enough in terms of accuracy and reliability, indicating a need for improvement. The objective of this study is to develop a construction cost budget prediction model that can be utilized in the early stages of public building projects using an artificial neural network (ANN). In this study, an artificial neural network model was developed using the SPSS Statistics program and the data provided by the Public Procurement Service. The level of construction cost budget prediction was analyzed, and the accuracy of the model was validated through additional testing. The validation results demonstrated that the developed artificial neural network model exhibited an error range for estimates that can be utilized in the early stages of projects, indicating the potential to predict construction cost budgets more accurately by incorporating various project conditions.

• Korea Trade Review
• /
• v.43 no.4
• /
• pp.155-175
• /
• 2018

This study estimates the technical efficiency and total factor productivity(TFP) of and analyzes the relationship between TFP and exports for Korean manufacturing companies from 2000 to 2016. Specially, TFP is decomposed into Technical Change(TC), Technical Efficiency Change (TEC), and Sale Effect(SE), and compared between large and small enterprises. First, in the case of technical efficiency, the Korean economy has been very vulnerable to external shocks, such as the sharp decline following the 2008 financial crisis. The efficiency of the electronics, automobile, and machinery sectors is low and needs to be improved. In addition, the technological efficiency of large enterprises is higher than that of SMEs in most manufacturing sub-sectors except for non-ferrous metals. In the case of TFP, most changes are due to TC, and the effective combination of labor, capital and the effect of scale have little effect, suggesting that improvement of internal structure is urgent. In addition, volatility due to the impact of the financial crisis in 2008 was much larger in SMEs than in large companies, so external economic impacts are more greater for SMEs than large enterprises. The relationship between TFP decomposition factors and exports shows that TC has a positive effect only on exports of SMEs. Therefore, in order to increase exports, in the case of SMEs, R&D support to promote technological development is needed. In the case of large companies, it is necessary to establish differentiated strategies for each export market, competitor company, and item to link efficiency and scale effect of exports.

• The Korean Journal of Applied Statistics
• /
• v.36 no.5
• /
• pp.381-397
• /
• 2023

The most representative design used in clinical trials is randomization, which is used to accurately estimate the treatment effect. However, comparison between the treatment group and the control group in an observational study without randomization is biased due to various unadjusted differences, such as characteristics between patients. Propensity score weighting is a widely used method to address these problems and to minimize bias by adjusting those confounding and assess treatment effects. Inverse probability weighting, the most popular method, assigns weights that are proportional to the inverse of the conditional probability of receiving a specific treatment assignment, given observed covariates. However, this method is often suffered by extreme propensity scores, resulting in biased estimates and excessive variance. Several alternative methods including trimming, overlap weights, and matching weights have been proposed to mitigate these issues. In this paper, we conduct a simulation study to compare performance of various propensity score weighting methods under diverse situation, such as limited overlap, misspecified propensity score, and treatment contrary to prediction. From the simulation results overlap weights and matching weights consistently outperform inverse probability weighting and trimming in terms of bias, root mean squared error and coverage probability.