• The Journal of the Korea Contents Association
• /
• v.21 no.10
• /
• pp.1-10
• /
• 2021

Log data accumulated in the Learning Management System (LMS) provide high-quality information for the learning process of students. Until now, various studies have been conducted to predict students' academic achievement using LMS log data. However, previous studies were based on relatively small sample sizes of students and courses, limiting the possibility of generalization. This study developed and validated a deep neural network model for the early prediction of academic achievement of college students using massive LMS log data. To this end, we used 78,466,385 cases of LMS log data and 165,846 cases of grade data. The proposed model predicted the excellent-grade students with a high level of accuracy from the beginning of the semester. Meanwhile, the prediction accuracy for the moderate and underachieving groups was relatively low, but the accuracy improved as the time points of the prediction were delayed. This study is meaningful in that we developed an early prediction model based on a deep neural network with sufficient accuracy for practical utilization by only using LMS log data.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.21 no.1
• /
• pp.47-56
• /
• 2015

Under the conditions of invested actual changes of heeling angles during MV Sewol's sinking, this study proposes passengers' evacuation scenarios, which are based on the assumption of normal orders of evacuation being given to the passengers, and evaluates using a marine-specialized human evacuation simulation tool. As results, when the heeling angle is set as 0 degree or 30 degree, it is found out that almost every passengers can success to evacuate to the musterstations, even though the evacuation times are different depending on the scenarios and the walking speeds. Meanwhile, when the heeling angle is varied as the Sewol incident, 3.1 %(Scenario Sc-Va which set chutes on port side as evacuation routes), 11.1 %(Sc-Vb, every open decks of port side as evacuation routes) and 20.0 %(Sc-Vc, every open decks of port and AFT sides as evacuation routes) among 476 passengers can successfully reach to the musterstations from their cabins with the condition of average walking speed as 2.04 m/s on flat. And only 0.8 %(Sc-Va), 3.8 %(Sc-Vb) and 10.7 %(Sc-Vc) can success to evacuate with the condition of average walking speed as 1.48 m/s on flat.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.44 no.5
• /
• pp.45-54
• /
• 2007

Time/frequency analysis has been extensively used in biomedical signal processing. By extracting some essential features from the electro-physiological signals, these methods are able to determine the clinical pathology mechanisms of some diseases. However, this method assumes that the signal should be stationary, which limits its application in non-stationary system. In this paper, we develop a new signal processing method using Hilbert-Huang Transform to perform analysis of the nonlinear and non-stationary ventricular fibrillation(VF). Hilbert-Huang Transform combines two major analytical theories: Empirical Mode Decomposition(EMD) and the Hilbert Transform. Hilbert-Huang Transform can be used to decompose natural data into independent Intrinsic Mode Functions using the theories of EMD. Furthermore, Hilbert-Huang Transform employs Hilbert Transform to determine instantaneous frequency and amplitude, and therefore can be used to accurately describe the local behavior of signals. This paper studied for Return Of Spontaneous Circulation(ROSC) and non-ROSC prediction performance by Support Vector Machine and three parameters(EMD-IF, EMD-FFT) extracted from ventricular fibrillation ECG waveform using Hilbert-Huang transform. On the average results of sensitivity and specificity were 87.35% and 76.88% respectively. Hilbert-Huang Transform shows that it enables us to predict the ROSC of VF more precisely.

• Journal of Korea Spatial Information System Society
• /
• v.7 no.1 s.13
• /
• pp.3-11
• /
• 2005

This paper proposes the method for the man power forecasting and performs mid-term(1994-1998) forecasting of telematics man power demands in Korea. Telematics technology has been selected as '839 New IT Growth Engine' by Ministry of Information and Communication (MIC) of Korean Government to boost Korean IT industry for the next 10 years. In order to meet the man power requirement in this telematics industry, accurate forecasting of the man power demand is necessary. The procedures for the forecasting includes study of man power forecasting models, deriving market size of the telematics industry, perform labor productivity analysis, derive the man power structure by the types of the work forces by the types of telematics industry, and finally derive annual man power demands by the worker types and the telematics industry types.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.3
• /
• pp.579-586
• /
• 2018

With the beginning of the new era of bigdata, information extraction or prediction are an important research area. Here, we present the acquisition of semi-automatically curated large-scale biological database and the prediction of enzyme reaction annotation for analyzing the pharmacological activities of drugs. Because the xenobiotic metabolism of pharmaceutical drugs by cellular enzymes is an important aspect of pharmacology and medicine. In this study, we apply and analyze similarity models to predict bimolecular reactions between human enzymes and their corresponding substrates. Thirteen models select to reflect the characteristics of each cluster in the similarity model. These models compare based on sensitivity and AUC. Among the evaluation models, the Simpson coefficient model showed the best performance in predicting the reactivity between the enzymes. The whole similarity model implement as a web service. The proposed model can respond dynamically to the addition of reaction information, which will contribute to the shortening of new drug development time and cost reduction.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.441-446
• /
• 2006

유역에서 발생한 고탁수가 저수지내에서 장기화하는 현상에 관한 대책은 장기간이 소요되는 유역관리 대책과 탁수저감일수를 최소화하는 선택취수 개념의 우선 방류기법 등이 있다. 하류하천으로의 탁수 우선배제를 실시하기 위해서는 댐 방류탁수가 하류에 미치는 영향을 정확히 예측할 수 있는 하천탁도 예측 및 관리시스템 구축이 필요하다. 최근 고탁수현상을 경험한 낙동강 수계에서는 본류의 탁도 관리를 위해 임하댐 및 안동댐의 연직 탁도 분포 변화에 따라 댐 방류탁수에 의한 영향을 최소 또는 기준치 이하로 할 수 있는 댐 연계운영에 관한 조건 도출이 필요하다. 선택취수에 관한 수치모의 결과를 이용하여 저수지내 온도 및 탁도 분포 변화와 취수탑의 수문운영에 따른 방류수의 온도 및 탁도를 예측할 수 있다. 본 연구에서는 이와 같은 결과를 이용하여 안동댐과 임하댐의 방류조건에 대한 하류하천 합류부에서의 2차원 이송 확산 수치모의를 수행하고자 하였다. 이와 같은 연구를 성공적으로 수행하기 위해서는 지속적인 현장조사를 통한 지점별 탁도-SS 상관관계 도출, 비정상 흐름 및 수질모의 검 보정, 탁도예측 결과 검증이 필요하다. 평면 2차원 흐름 및 수질 모의 결과에 의하면, 임하댐 방류 탁도로 인한 반변천의 고탁수는 안동댐 방류 수에 지배를 받는 낙동강 본류와 합류되는 지점에서부터 약 5 km 떨어진 지점에서 횡방향 완전혼합이 이루어지는 것으로 나타났다. 이와 같은 모의결과는 완전혼합을 가정하는 1차원 수질모델링의 초기 입력자료에 사용됨으로써 1차원 탁도 모의결과의 정확성을 높이는 데 사용될 수 있다. 이를 이용하여 낙동강 수계 댐 연계운영에 따른 낙동강 탁도 예측모의를 수행하고, 이 결과를 이용하여 낙동강 탁도 예경보 시스템을 구축해야 할 것이다. 또한 선택취수 등을 통해 저수지 관리를 효과적으로 수행하기 위해서는 저수지 내부의 탁도 거동을 정확히 예측할 수 있어야 한다. 따라서 추후 동수역학 및 열역학에 기초한 3차원 수치모형 연구와 성층흐름에 정밀한 밀도류 실험연구 및 이에 대한 적용이 필요할 것으로 판단된다.

• Journal of Distribution Research
• /
• v.10 no.3
• /
• pp.1-14
• /
• 2005

It is important to control performance and a Sustainable Collaboration (SC) for the successful Supply Chain Management (SCM). This research developed a control model which analyzed SCM performances based on a Balanced Scorecard (ESC) and an SC using Support Vector Machine (SVM). 108 specialists of an SCM completed the questionnaires. We analyzed experimental data set using SVM. This research compared the forecasting accuracy of an SCMSC through four types of SVM kernels: (1) linear, (2) polynomial (3) Radial Basis Function (REF), and (4) sigmoid kernel (linear > RBF > Sigmoid > Polynomial). Then, this study compares the prediction performance of SVM linear kernel with Artificial Neural Network. (ANN). The research findings show that using SVM linear kernel to forecast an SCMSC is the most outstanding. Thus SVM linear kernel provides a promising alternative to an SC control level. A company which pursues an SCM can use the information of an SC in the SVM model.

• Korean Journal of Construction Engineering and Management
• /
• v.12 no.1
• /
• pp.62-72
• /
• 2011

Currently, The failure of construction project is increasing to be caused by a changing construction environment. According to this circumstances, Researches of project success factors affecting performance have been presented to develop strategies for efficient construction project management in the construction industry. However, Conducting efficient construction project management is difficult because project manager could not know which project success factors can be improved or not. Especially, although the project characteristics were derived the level of difficulty for performance, research of the project characteristics which could not be improved as influence factor to performance is lacking. Therefore, This paper has developed the Performance Predicting System(PPS) with Fuzzy set theory to establish. PPS has been developed to establish efficient project management strategies and to save time and effort. As Contractor inputs the project characteristics, PPS can predict the level of difficulty of performance.

• Clean Technology
• /
• v.18 no.3
• /
• pp.265-271
• /
• 2012

The use of renewable energies has been required to diminish the dependency on fossil fuels. As one of clean energy sources biomass has been extensively studied because various biomass resources necessitated rapid characterization of their chemical and physical properties in an on-line or real-time basis. For such an analysis near-infrared (NIR) spectroscopy has been successfully applied because of its non-invasive and informative characteristics. In this work, the applicability of nonlinear chemometric techniques based on biomass near infrared (NIR) data is evaluated for the rapid prediction of ash/char contents in different types of biomass. The prediction results of various prediction models and the effect of using preprocessing methods for NIR data are compared using six types of biomass NIR data. The results showed that nonlinear prediction models yielded better prediction performance than linear ones. It also turned out that by adopting the use of proper preprocessing methods the performance of prediction of biomass properties improved.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.11
• /
• pp.1608-1614
• /
• 2022

As competition between universities intensifies due to the recent decrease in the number of students, it is recognized as an essential task of universities to predict students who are underperforming at an early stage and to make various efforts to prevent dropouts. For this, a high-performance model that accurately predicts student performance is essential. This paper proposes a method to improve prediction performance by removing or amplifying abnormal data in a classification prediction model for identifying underperforming students. Existing anomaly data processing methods have mainly focused on deleting or ignoring data, but this paper presents a criterion to distinguish noise from change indicators, and contributes to improving the performance of predictive models by deleting or amplifying data. In an experiment using open learning performance data for verification of the proposed method, we found a number of cases in which the proposed method can improve classification performance compared to the existing method.