• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.33 no.1
• /
• pp.287-295
• /
• 2013

This paper developed advanced gravity model for higher estimation accuracy, that deals with residuals. Previously studied paper using gravity model's residual, residual calculated that observed O-D value minus derived O-D value from gravity model, and this residual added to the target year's estimated value from gravity model. In this paper, residuals calculated on gravity model parameter estimation process, and this residual is revealed the same value that observed O-D value devided by derived O-D value from gravity model. And case study resulted that developed new gravity model that applied accordance rate of observed O-D value and derived O-D value from gravity model has higher estimation accuracy than other gravity models as basic gravity model and residual plused gravity model.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.3
• /
• pp.176-182
• /
• 2024

In this study, we compared and analyzed various methods of missing data handling to build a machine learning model that can effectively analyze and predict time series data with a high percentage of missing values. For this purpose, Predictive State Model Filtering (PSMF), MissForest, and Imputation By Feature Importance (IBFI) methods were applied, and their prediction performance was evaluated using LightGBM, XGBoost, and Explainable Boosting Machines (EBM) machine learning models. The results of the study showed that MissForest and IBFI performed the best among the methods for handling missing values, reflecting the nonlinear data patterns, and that XGBoost and EBM models performed better than LightGBM. This study emphasizes the importance of combining nonlinear imputation methods and machine learning models in the analysis and prediction of time series data with a high percentage of missing values, and provides a practical methodology.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2008.10a
• /
• pp.324-328
• /
• 2008

금융산업에서 고객의 이탈비율은 기대수익에 영향을 미친다는 점에서 예측이 필요한 부분이며 최근 들어 정확한 예측을 통한 비용관리가 이루어지면서 고객 이탈을 예측하는 것이 중요한 문제로 떠오르고 있다. 그러나 보험 고객 데이터가 대용량이고 불균형한 출력 값을 갖는 특성으로 인해 기존의 방법으로 예측 모델을 만드는 것이 적합하지 않다. 본 연구에서는 대용량 데이터를 처리하는 데 효과적으로 알려져 있는 Trust-region Newton method를 적용한 로지스틱 회귀분석을 통해 이탈고객을 예측하는 것을 주된 연구로 하며, 불균형한 데이터에서의 예측정확도를 높이기 위해 Oversampling, Clustering, Boosting 등을 이용하여 고객 데이터에 적합한 이탈 고객 예측 모형을 제시하고자 한다.

• Journal of Korean Society of Forest Science
• /
• v.106 no.4
• /
• pp.480-486
• /
• 2017

This study was conducted to develop the merchantable stem volume models to predict the volume up to upper diameter or upper height out of the total stem volume, targeting on Pinus densiflora, Pinus koraiensis, and Larix kaempferi in South Korea. The 131 stemmed sample trees for stem analysis were used as the data for developing the models. The six kinds of merchantable volume equations including merchantable volume ratio form, ratio form, and exponential ratio form were examined to develop the best models. The two models were finally selected as the best models to predict the merchantable volume: $V_d=V_t\{{\exp}[{\alpha}_1(d^{{\alpha}_2}/D^{{\alpha}_3})]\}$ for upper diameter and $V_h=V_t\{1+{\beta}_1(P^{{\beta}_2}/H^{{\beta}_3})\}$ for upper height. By rearranging the best model equations, implicit taper functions were derived, and the estimation was performed for the upper height by upper diameter and upper diameter by upper height. Because of not only the high accuracy but also the convenience, the models developed in this study were considered to be easily applicable in the field of forestry.

• Korean Journal of Clinical Laboratory Science
• /
• v.50 no.3
• /
• pp.245-252
• /
• 2018

We evaluated the possible clinical application of Non HDL-cholesterol and triglyceride to HDL-cholesterol ratio as a metabolic syndrome predictor for the elderly in Korea. 1,543 elderly persons aged 65 years or older who visited the health examination center of Gyeonggi Regional General Hospital from January 2015 to December 2017 and had a health checkup were enrolled in this study. Metabolic syndrome was diagnosed based on the American Heart Association/National Heart, Lung, and Blood Institute (AHA/NHLBI) standards. Abdominal obesity was assessed by the Asia-Pacific standards presented at the World Health Organization (WHO) West Pacific Region. Non-HDL-cholesterol was calculated as the difference between total cholesterol and HDL-cholesterol. The metabolic syndrome predictive power was higher for triglyceride to HDL-cholesterol ratio than for Non HDL-cholesterol. After correcting for related factors, triglyceride to HDL-cholesterol ratio was higher in the $4^{th}$ quartile, which had a higher risk of developing metabolic syndrome, than in the $1^{st}$ quartile. The optimal cutoff value for the triglyceride to HDL-cholesterol ratio that predicts the onset of metabolic syndrome was 2.8. triglyceride to HDL-cholesterol ratio can be a simple and practical indicator of the risk of metabolic syndrome.

• Korean Journal of Clinical Laboratory Science
• /
• v.51 no.1
• /
• pp.42-49
• /
• 2019

This study compared the utility of the serum uric acid/creatinine ratio with that of uric acid as a risk predictor of metabolic syndrome. From November 2016 to October 2018, 14,190 adult men under the age of 20 years, who underwent a comprehensive health checkup at a general hospital in their metropolitan area, were included. Metabolic syndrome was assessed according to the American Heart Association/National Heart Lung and Blood Institute (AHA/NHLBI) criteria. Abdominal obesity was based on the WHO criteria in the Western Pacific region. The serum uric acid/creatinine ratio was found to be higher in the fourth quartile than in the first quartile, with a high incidence of metabolic syndrome and metabolic syndrome components. On the other hand, ROC analysis revealed the serum uric acid/creatinine ratio to be a similar indicator of the metabolic syndrome risk to serum uric acid (AUC, 0.554 vs 0.566). The serum uric acid/creatinine ratio showed lower sensitivity and higher specificity than uric acid. In conclusion, the utility of the serum uric acid/creatinine ratio as an independent indicator to predict the risk of metabolic syndrome is limited, and should be used only as an auxiliary marker.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10c
• /
• pp.682-684
• /
• 2002

최근 여러 논문에서 실 데이터 종속을 제거하기 위하여 결과 값 예상 기법을 제안하였다. 결과 값 예상 기법 중 혼합형 결과 값 예측기는 다양한 패턴을 갖는 명령어를 모두 예측함으로써 높은 예상 정확도를 얻을 수 있지만 하나의 명령어가 여러 개의 예측기 테이블에 중복 저장되어 높은 하드웨어 비용을 요구한다는 단점이 있다. 본 논문에서는 이러한 단점을 극복하기 위하여 프로파일링으로 얻어진 정적 분류 정보를 사용하여, 명령어률 예상 정확도가 높은 예측기에만 할당하여 예상 테이블 크기를 감소 시켰다. 또한 동적으로 적절한 예측기를 선택하도록 함으로써 예상 정확도를 더욱 향상 시켰다. 본 논문에서는 SPECint95 벤치마크 프로그램에 대해 SimpleScalar/PISA 3.0 툴셋을 사용하여 실험하였다. 정적-동적 분류 정보를 모두 사용하였을 경우 87.9%, VHT 크기를 4K로 축소한 경우 87.5%로 비슷한 예상정확도를 얻으면서 예상 테이블의 크기는 50%로 감소하였다. 또한 실행 패턴의 유형 비율에 따라 각 예측기의 VHT를 구성한 경우 예상 테이블 크기를 25%로 줄일 수 있었다.

• Proceedings of the KAIS Fall Conference
• /
• 2009.12a
• /
• pp.8-11
• /
• 2009

성공적인 창업을 위해서는 창업가의 준비가 선행되어야 하지만, 매력적인 사업 아이디어의 계발이 뒤따라야 한다. 그러나 지금까지의 창업연구는 창업행동과 사업성과에 영향을 미치는 창업가 요인에 치우쳐 있으며, 사업 아이디어의 계발과 평가에 대한 연구는 부족한 실정이다. 이 연구는 고객이 상품을 구매하기 전 인터넷 검색엔진에서 해당 상품에 대한 검색을 하는 경우가 일반화되고 있다는 사실과 고객이 검색엔진에 입력하는 키워드는 고객의 의도를 대변한다는 사실을 기초로, 키워드로 표현된 사업 아이디어의 매력도를 객관적으로 측정하는 방법을 제시하는 것을 목적으로 한다.이 연구는 키워드로 표현된 사업 아이디어 매력도(BIA)를 구매의도를 가진 잠재고객의 자사 웹 사이트 방문수로 정의한다. 키워드로 표현된 사업 아이디어 매력도(BIA)는 [해당 키워드의 조회수(Q) ${\times}$ 구매의도 비율(R) / 경쟁 사이트의 수(S)]의 수식으로 나타낼 수 있으며, 수식을 구성하는 변수 중에서 해당 키워드의 조회수(Q)와 경쟁 사이트의 수(S)는 검색엔진에서 쉽게 제공 받을 수 있으므로, 구매의도 비율(R)만 알 수 있다면 BIA를 비교적 정확히 추정할 수 있다. 연구자는 특정 분야 키워드 100개를 선정한 다음, 전문가로 하여금 각 키워드의 구매의도 비율(R)을 추정하게 하고, 전문가 추정 없이도 구매의도 비율을 예측할 수 있도록 각 키워드의 구매의도 비율(R)을 예측하는 주요 데이터를 의사결정 나무 기법으로 도출하고, 의사결정 나무 기법으로 도출된 데이터로 구성된 회귀식을 제시함으로써 키워드로 표현된 사업 아이디어 매력도(BIA)를 객관적으로 평가하는 방법을 제시한다. 이 연구는 사업 아이디어의 계발과 평가에 대한 객관적인 기준을 제시함으로써 창업의 성공률을 높이는 데 기여할 수 있고, 창업연구에 새로운 방법론을 도입했다는 점에서 의의가있다.

• The Korean Journal of Financial Management
• /
• v.23 no.2
• /
• pp.209-237
• /
• 2006

Since previous default forecasting models for the firms evaluate the probability of default based upon the accounting data from book values, they cannot reflect the changes in markets sensitively and they seem to lack theoretical background. The market-information based models, however, not only make use of market data for the default prediction, but also have strong theoretical background like Black-Scholes (1973) option theory. So, many firms recently use such market based model as KMV to forecast their default probabilities and to manage their credit risks. Korean firms also widely use the KMV model in which default point is defined by liquid debt plus 50% of fixed debt. Since the debt structures between Korean and American firms are significantly different, Korean firms should carefully use KMV model. In this study, we empirically investigate the importance of debt structure. In particular, we find the following facts: First, in Korea, fixed debts are more important than liquid debts in accurate prediction of default. Second, the percentage of fixed debt must be less than 20% when default point is calculated for Korean firms, which is different from the KMV. These facts give Korean firms some valuable implication about default forecasting and management of credit risk.

• Journal of Advanced Navigation Technology
• /
• v.19 no.3
• /
• pp.199-206
• /
• 2015

For real-time precise positioning, IGS provides ephemeris predictions (IGS ultra-rapid, IGU) and real-time ephemeris estimates (real-time service, RTS). Due to the RTS data latency, which ranges from 5 s to 30 s, a short-term prediction process is necessary before applying the RTS corrections. In this paper, the real-time performance of the RTS correction and IGU prediction are compared. The RTS correction availability for the GPS satellites observed in Korea is computed as 99.3%. The RTS correction is applied to broadcast ephemeris to verify the accuracy of the RTS correction. The 3D orbit RMS error of the RTS correction is 0.043 m. Prediction of the RTS correction is modeled as a polynomial, and then the predicted value is compared with the IGU prediction value. The RTS orbit prediction accuracy is nearly equivalent to the IGU prediction, but RTS clock prediction performance is 0.13 m better than the IGU prediction.