• Journal of Digital Contents Society
• /
• v.19 no.1
• /
• pp.141-147
• /
• 2018

This study analyzed the early retired employees who stayed in company no longer than 3 years based on a certain company's personality evaluation result data. The predicted model was analyzed by dividing into two categories; the manufacture group and the R&D group. Independent variables were selected according to the stepwise method. A logistic regression model was selected as a prediction model among various supervised learning methods, and trained through cross-validation to prevent over-fitting or under-fitting. The accuracy of the two groups were confirmed by the confusion matrix. The most influential factor for early retirement in the manufacture group was revealed as "immersion," and for the R&D group appeared as "antisocial." In the past, people concentrated on collecting data by questionnaire and identifying factors that are highly related to the retirement, but this study suggests a sustainable early retirement prediction model in the future by analyzing the tangible outcome of the recruitment process.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.06a
• /
• pp.359-368
• /
• 2001

본 연구에서는 부도예측용 인공신경망의 입력노드 선정을 위한 휴리스틱으로 연결강도분석 접근법을 제안한다. 연결강도분석은 학습이 끝난 인공신경망에서 입력노드와 은닉노드와 연결된 가중치의 절대값 즉, 연결강도를 분석하여 입력변수를 선정하는 접근법으로, 본 연구에서는 약체연결뉴론제거법, 강체연결뉴론선택법 그리고 이 두 기법을 통합한 통합 연결강도 모형을 제안하여 각각 의사결정 트리 및 다변량판별분석에 의해 선정된 입력변수를 이용한 인공신경망 모형과 예측율을 비교한다. 실험 결과 본 연구에서 제안하고 있는 방법론이 의사결정트리나 다다변량판별분석 기법 보다 높은 예측율을 보여 주었다. 특히 두 기법의 통합연결강도 모형의 경우에는 다른 단일 기법보다 높은 예측율을 보이고 있다.

• Journal of Science Education
• /
• v.46 no.1
• /
• pp.80-92
• /
• 2022

We fitted latent growth models of attitude towards science using the Korea Education & Employment Panel 2004-2007 data with 343 high school students. The growth model show better fit indices compared to the no growth model. The intercept and slope showed significant variances, and thus, we added control variables of the number, ratio of advanced courses, and variety in science elective subjects, and the achievement percentile for middle school. In the conditional growth model, the previous achievement has significant positive effects on the intercept and the ratio of the advanced courses and variety of science subjects show significantly positive effects on the slope. Based on the results, it supports the 2022 Revised Science Curricular that high school credit system should provide students with basic 'Physics,' 'Chemistry,' 'Biology,' and 'Earth Science,' credits in 'general electives', various integrated subjects in 'converged electives', and highly advanced subjects in 'career electives.'

• Proceedings of the Korea Water Resources Association Conference
• /
• 2011.05a
• /
• pp.257-257
• /
• 2011

현업에서 사용하는 유출해석 기본이론은 연속방정식과 운동방정식으로서 운동파가정(kinematic wave analogy)을 기반으로 한 집중수문모형(lumped hydrologic model)에 의하여 수행되고 있지만 집중형 모형은 한 매개변수에 여러 가지의 물리적 과정을 개념화하여 담고 있기 때문에 유출과정에 대한 섬세한 모형화의 제약으로 인하여 유역고유의 매개변수값을 찾기가 쉽지 않은 단점을 가지고 있다. 이에 본 연구에서는 물리적 기반의 2차원 분포형 강우-유출모형을 개발하고자 하며 이는 완전분포형 수문동역학적 모형으로 지표흐름과 침투과정, 기저유출과 관련된 과정을 모의한다. 본 모형은 공간적으로 변화하는 침투량과 소규모 및 대규모의 지형학적 특성을 사용하는 St. Venant 방정식을 사용하고 개발될 모형은 모든 스케일에서의 수심과 유량을 계산할 수 있으며 Richard 방정식(또는 선택적으로 Green-Ampt 방정식 채택)을 이용하여 정밀한 침투량 모의가 가능하다. 또한 레이다등의 고해상도 강우관측자료를 지점자료와 합성하여 입력자료로 사용할 수 있도록하고자 하며 강우-유출모형에 다목적댐이나 보등에서의 유량조절효과를 반영하고, 다목적댐군에서의 연계운영모의가 가능케 함으로서 현업의 운영자들이 실무에서 실질적으로 활용할 수 있는 형태의 모형을 개발하고자 한다. 이는 국내에서의 2차원 분포형 강우-유출모형을 자체 개발함으로서 연구역량을 제고하고, 국내 현업기관에서의 분포형 모형기반의 홍수모니터링 및 전망시스템의 확산에 기여할 것으로 예상된다.

• Proceedings of the KOR-KST Conference
• /
• 1998.10a
• /
• pp.496-505
• /
• 1998

물류활동에서 운송부분이 차지하는 중요성은 기업의 경영자층에게 더욱 중요하게 인식되고 있는데, 이는 경쟁 환경 때문이다. 경쟁환경에서는 'Output Logistic' 즉, 운송활동의 수행정도에 따라 물류활동의 성패가 결정된다. 따라서 기업물류활동의 근간인 운송활동의 주요 결정요인과 선택형태를 알아보고, 이들이 실제로 기업의 물류활동에 어느 정도 영향을 미치는지를 심층적으로 분석해보는 것도 매우 의미 있는 일이라 하겠다. 더구나 우리 나라의 현재의 경제여건에서는 물류비에 대한 효율화 작업이 필요한데 반해 그 동안의 연구들을 살펴보면 SP 자료를 이용한 가상적 상황하에서의 화주의 선택행태를 분석하였으므로 실제 선택한 수단간의 gap을 극복할 수 없었다. 기업은 운송수단의 선택시 복잡한 결정과정을 갖는다. 이는 운송부문이 총물류비용에서 차지하는 중요도 때문이다. 기업의 운송관리자는 화물을 출하할 때마다 선택의 기로에 서게 된다. 즉, 일부는 조직 체계나 다른 계약 여건에 따라 이전과 동일한 수단을 선택하는 경우도 있지만, 많은 경우에는 매번 출하시 마다 최적의 운송수단을 선택하기 위한 새로운 결정을 하게 된다. 본 연구는 이러한 화주의 수단선택행태를 실제 RP 자료를 이용하여 분석하였다. 수단선택모형의 적용 및 분서결과를 살펴보면 상당히 attractive한 결과를 발견할 수 있는데 각 품목별 추정 값이 운송거리에 대해서는 음으로 운송비용에 대해서는 양으로 나타나고 있다. 다시 말하면 운송거리가 길수록 효용은 감소하고 운송비용이 커질수록 효용은 증가한다는 것을 의미하므로 그 분석결과가 올바른 결과를 도출하고 있지는 않다. 그러나 여기서 알수 있는 것은 운송거리와 운송비용이 각각 주요한 변수라는 것이다. 모형의 타당성을 검증하기 위해서는 logilikelihood 값을 구하여 $\rho$^2분석을 시행하였다. 여기서는 각 품목별로 $\rho$^2값이 약 0.15~0.3의 비교적 높은 수치를 보여주고 있으므로 모형의 설명력이 어느 정도 있다는 것이 아울러 증명이 되었다. 상관관계에 대한 분석에서는 영업용 차량간의 상관관계가 높게 나타났으며, 이는 곧 영업용 화물차량을 적재중량별로 구분하는 것이 별 의미가 없음을 의미한다. 다시 말하면 자가용 차량을 보유하고 있지 않은 회사는 다른 운송전문업체에 화물운송을 의뢰하게 되므로 출하중량에 따라 화물차량을 구분하는 것에 대해서 그다지 큰 고려를 하지 않는 것으로 해석할 수가 있다.

• Journal of Intelligence and Information Systems
• /
• v.22 no.1
• /
• pp.139-157
• /
• 2016

Bankruptcy involves considerable costs, so it can have significant effects on a country's economy. Thus, bankruptcy prediction is an important issue. Over the past several decades, many researchers have addressed topics associated with bankruptcy prediction. Early research on bankruptcy prediction employed conventional statistical methods such as univariate analysis, discriminant analysis, multiple regression, and logistic regression. Later on, many studies began utilizing artificial intelligence techniques such as inductive learning, neural networks, and case-based reasoning. Currently, ensemble models are being utilized to enhance the accuracy of bankruptcy prediction. Ensemble classification involves combining multiple classifiers to obtain more accurate predictions than those obtained using individual models. Ensemble learning techniques are known to be very useful for improving the generalization ability of the classifier. Base classifiers in the ensemble must be as accurate and diverse as possible in order to enhance the generalization ability of an ensemble model. Commonly used methods for constructing ensemble classifiers include bagging, boosting, and random subspace. The random subspace method selects a random feature subset for each classifier from the original feature space to diversify the base classifiers of an ensemble. Each ensemble member is trained by a randomly chosen feature subspace from the original feature set, and predictions from each ensemble member are combined by an aggregation method. The k-nearest neighbors (KNN) classifier is robust with respect to variations in the dataset but is very sensitive to changes in the feature space. For this reason, KNN is a good classifier for the random subspace method. The KNN random subspace ensemble model has been shown to be very effective for improving an individual KNN model. The k parameter of KNN base classifiers and selected feature subsets for base classifiers play an important role in determining the performance of the KNN ensemble model. However, few studies have focused on optimizing the k parameter and feature subsets of base classifiers in the ensemble. This study proposed a new ensemble method that improves upon the performance KNN ensemble model by optimizing both k parameters and feature subsets of base classifiers. A genetic algorithm was used to optimize the KNN ensemble model and improve the prediction accuracy of the ensemble model. The proposed model was applied to a bankruptcy prediction problem by using a real dataset from Korean companies. The research data included 1800 externally non-audited firms that filed for bankruptcy (900 cases) or non-bankruptcy (900 cases). Initially, the dataset consisted of 134 financial ratios. Prior to the experiments, 75 financial ratios were selected based on an independent sample t-test of each financial ratio as an input variable and bankruptcy or non-bankruptcy as an output variable. Of these, 24 financial ratios were selected by using a logistic regression backward feature selection method. The complete dataset was separated into two parts: training and validation. The training dataset was further divided into two portions: one for the training model and the other to avoid overfitting. The prediction accuracy against this dataset was used to determine the fitness value in order to avoid overfitting. The validation dataset was used to evaluate the effectiveness of the final model. A 10-fold cross-validation was implemented to compare the performances of the proposed model and other models. To evaluate the effectiveness of the proposed model, the classification accuracy of the proposed model was compared with that of other models. The Q-statistic values and average classification accuracies of base classifiers were investigated. The experimental results showed that the proposed model outperformed other models, such as the single model and random subspace ensemble model.

• Journal of Korea Water Resources Association
• /
• v.51 no.8
• /
• pp.677-686
• /
• 2018

There remain numerous ungauged watersheds in Korea owing to limited spatial and temporal streamflow data with which to estimate hydrological model parameters. To deal with this problem, various regionalization approaches have been proposed over the last several decades. However, the results of the regionalization models differ according to climatic conditions and regional physical characteristics, and the results of the regionalization models in previous studies are generally inconclusive. Thus, to improve the performance of the regionalization methods, this study attaches hydrological model parameters obtained using a spatial proximity model to the explanatory variables of a regional regression model and defines it as a hybrid regionalization model (hybrid model). The performance results of the hybrid model are compared with those of existing methods for 37 test watersheds in South Korea. The GR4J model parameters in the gauged watersheds are estimated using a shuffled complex evolution algorithm. The variation inflation factor is used to consider the multicollinearity of watershed characteristics, and then stepwise regression is performed to select the optimum explanatory variables for the regression model. Analysis of the results reveals that the highest modeling accuracy is achieved using the hybrid model on RMSE overall the test watersheds. Consequently, it can be concluded that the hybrid model can be used as an alternative approach for modeling ungauged watersheds.

• Spatial Information Research
• /
• v.10 no.2
• /
• pp.275-287
• /
• 2002

Rainfall-Runoff Analysis in Whangryong River Basin was made using HEC-HMS and HEC-GeoHMS. The Basin was divided into three sub-basins using HEC-CeoHMS and GIS. Then, GIS input data were derived from each sub-basins. SCS CN runoff-volume model, Snyder's UH direct-runoff model, exponential recession baseflow model and Muskingum routing model in HEC-HMS were used to simulate the runoff volume using selected rainfall event and the parameters were optimized. Peak flowrate calculated using optimized parameters was compared to the observed flowrate in the basin. The result proved to be good agreement with each other. Optimized parameters in this local basin can be used to calculate the peak flowrate in the future.

• Journal of Korean Society of Transportation
• /
• v.27 no.6
• /
• pp.157-166
• /
• 2009

A network design problem (NDP) is to find a design parameter to optimize the performance of transportation system. This paper presents a modified NDP, called target-oriented NDP, which contains a target that we try to arrive in real world, and also proposes a solution algorithm. Unlike general NDP which seeks an optimal value to minimize or to maximize objective function of the system, in target-oriented NDP traffic manager or operator can set a target level prior and then try to find an optimal design variable to attain this goal. A simple example for mode choice problem is given to test the model.

• Journal of Korean Society of Transportation
• /
• v.26 no.2
• /
• pp.89-100
• /
• 2008

In an urban network controlled by traffic signals, there is an interaction between the signal timing and the routes chosen by individual road users. This study develops a bi level programming model for traffic signal optimization in networks with path based traffic assignment. In the bi level programming model, genetic algorithm approach has been proposed to solve upper level problem for a signalized road network. Path based traffic assignment using column generation technique which is proposed by M.H. Xu, is applied at the lower-level. Genetic Algorithm provieds a feasible set of signal timings within specified lower and upper bounds signal timing variables and feeds into lower level problem. The performance of this model is investigated in numerical experiment in a sample network. In result, optimal signal settings and user equilibrium flows are made.