• 대한산업공학회지
• /
• 제37권1호
• /
• pp.64-73
• /
• 2011

Oil price prediction is an important issue for the regulators of the government and the related industries. When employing the time series techniques for prediction, however, it becomes difficult and challenging since the behavior of the series of oil prices is dominated by quantitatively unexplained irregular external factors, e.g., supply- or demand-side shocks, political conflicts specific to events in the Middle East, and direct or indirect influences from other global economical indices, etc. Identifying and quantifying the relationship between oil price and those external factors may provide more relevant prediction than attempting to unclose the underlying structure of the series itself. Technically, this implies the prediction is to be based on the vectoral data on the degrees of the relationship rather than the series data. This paper proposes a novel method for time series prediction of using Semi-Supervised Learning that was originally designed only for the vector types of data. First, several time series of oil prices and other economical indices are transformed into the multiple dimensional vectors by the various types of technical indicators and the diverse combination of the indicator-specific hyper-parameters. Then, to avoid the curse of dimensionality and redundancy among the dimensions, the wellknown feature extraction techniques, PCA and NLPCA, are employed. With the extracted features, a timepointspecific similarity matrix of oil prices and other economical indices is built and finally, Semi-Supervised Learning generates one-timepoint-ahead prediction. The series of crude oil prices of West Texas Intermediate (WTI) was used to verify the proposed method, and the experiments showed promising results : 0.86 of the average AUC.

• 대한원격탐사학회지
• /
• 제39권5_3호
• /
• pp.891-907
• /
• 2023

Compact Advanced Satellite 500-4 (CAS500-4) is scheduled to be launched to collect high spatial resolution data focusing on vegetation applications. To achieve this goal, accurate surface reflectance retrieval through atmospheric correction is crucial. Therefore, a machine learning-based atmospheric correction algorithm was developed to simulate atmospheric correction from a radiative transfer model using Sentinel-2 data that have similarspectral characteristics as CAS500-4. The algorithm was then evaluated mainly for forest areas. Utilizing the atmospheric correction parameters extracted from Sentinel-2 and GEOKOMPSAT-2A (GK-2A), the atmospheric correction algorithm was developed based on Random Forest and Light Gradient Boosting Machine (LGBM). Between the two machine learning techniques, LGBM performed better when considering both accuracy and efficiency. Except for one station, the results had a correlation coefficient of more than 0.91 and well-reflected temporal variations of the Normalized Difference Vegetation Index (i.e., vegetation phenology). GK-2A provides Aerosol Optical Depth (AOD) and water vapor, which are essential parameters for atmospheric correction, but additional processing should be required in the future to mitigate the problem caused by their many missing values. This study provided the basis for the atmospheric correction of CAS500-4 by developing a machine learning-based atmospheric correction simulation algorithm.

• 융합정보논문지
• /
• 제8권3호
• /
• pp.175-181
• /
• 2018

본 연구의 목적은 학습자 역량을 선별, 그 역량에 대한 중요도-수행도 인식 차이를 알아보고, 교육요구도를 분석하여 학습지원 프로그램의 방향성을 제고하고자 함이다. 설문도구를 활용하여 온라인 방식(이메일 발송)과 오프라인 방식(서면 설문)을 병행하여 수합된 159부를 최종 분석에 사용하였다. 기술통계, 대응표본 t-검정(paired t-test), Borich 공식을 활용한 교육요구도 분석을 실시하였다. 연구결과, 1) 3개 영역, 10개 역량별 역량 중요도와 역량 수행도 인식 차이에 있어 모두 유의미한 결과를 확인하였고, 2) Borich 공식을 활용하여 10개 역량별 전체 교육요구도를 살펴본 결과, '전공분야 지식'이 1순위, '창의성'이 2순위, '문제해결력'이 3순위, '글로벌 역량'이 4순위, '테크놀로지 역량'이 5순위 등의 순으로 나타났다. 본 연구결과는 대학생 학습지원 프로그램을 기획 설계하는 측면에서 고려해야 할 학습자의 역량과 교육요구도에 따른 주제의 우선순위를 결정하고, 실행 평가하는 측면에서 프로그램 효과성을 판단하는데 실제적 준거로 활용될 수 있을 것으로 기대한다.

• 제어로봇시스템학회논문지
• /
• 제5권1호
• /
• pp.95-104
• /
• 1999

In this paper we propose a fast and precise control algorithm for a mobile robot, which aims at the self-tuning control applying two multi-layered neural networks to the structure of computed torque method. Through this algorithm, the nonlinear terms of external disturbance caused by variable task environments and dynamic model errors are estimated and compensated in real time by a long term neural network which has long learning period to extract the non-linearity globally. A short term neural network which has short teaming period is also used for determining optimal gains of PID compensator in order to come over the high frequency disturbance which is not known a priori, as well as to maintain the stability. To justify the global effectiveness of this algorithm where each of the long term and short term neural networks has its own functions, simulations are peformed. This algorithm can also be utilized to come over the serious shortcoming of neural networks, i.e., inefficiency in real time.

• 영어어문교육
• /
• 제15권2호
• /
• pp.209-224
• /
• 2009

This study examined the role of contextual knowledge in cross-cultural communication between non-native speakers on an interactive web with a bulletin board system through which college students of English at Japanese and Korean universities interacted with each other discussing the topics of local and global issues. The study investigated the influence of students' relative contextual knowledge on active participation in interactions and discussed the results focusing on the use of discourse strategies for meaning negotiation. The study argues that in interactions even between non-native speakers with limited proficiency, contextual knowledge in the topic under discussion affects the degree to which they accommodate to each other during communication and suggests that the focus of teaching English as a foreign language also should be given to what kind of contextual knowledge students need to obtain and how to express it rather than what level of proficiency in English they need to acquire.

• 한국과학교육학회지
• /
• 제16권4호
• /
• pp.461-469
• /
• 1996

The purpose of this study was to assess the attitudes of secondary school students in Korea toward science-technology related issues. A questionnaire was designed in which students were confronted with personal, global, and philosophical levels of arguments, which were composed of three against and three in favor of each eight issues, i.e., use of antibiotics, family planning, transplant of organs, genetic engineering, use of microorganisms. exploitation of the sea, land reclamation from the sea, and nature reserves. Student was requested to rate each argument independently and to vote for or against each issue. It was shown that most of students voted in favor of using technologies except land reclamation from the sea, and that students having more learning experiences on each topic vote more favorably. It is thought that our science education might be effective in increasing awareness and appreciation of benifits of technology, but it is not as effective in developing ambivalence attitudes.

• International Journal of Computer Science & Network Security
• /
• 제23권8호
• /
• pp.177-189
• /
• 2023

Malware detection is an increasingly important operational focus in cyber security, particularly given the fast pace of such threats (e.g., new malware variants introduced every day). There has been great interest in exploring the use of machine learning techniques in automating and enhancing the effectiveness of malware detection and analysis. In this paper, we present a deep recurrent neural network solution as a stacked Long Short-Term Memory (LSTM) with a pre-training as a regularization method to avoid random network initialization. In our proposal, we use global and short dependencies of the inputs. With pre-training, we avoid random initialization and are able to improve the accuracy and robustness of malware threat hunting. The proposed method speeds up the convergence (in comparison to stacked LSTM) by reducing the length of malware OpCode or bytecode sequences. Hence, the complexity of our final method is reduced. This leads to better accuracy, higher Mattews Correlation Coefficients (MCC), and Area Under the Curve (AUC) in comparison to a standard LSTM with similar detection time. Our proposed method can be applied in real-time malware threat hunting, particularly for safety critical systems such as eHealth or Internet of Military of Things where poor convergence of the model could lead to catastrophic consequences. We evaluate the effectiveness of our proposed method on Windows, Ransomware, Internet of Things (IoT), and Android malware datasets using both static and dynamic analysis. For the IoT malware detection, we also present a comparative summary of the performance on an IoT-specific dataset of our proposed method and the standard stacked LSTM method. More specifically, of our proposed method achieves an accuracy of 99.1% in detecting IoT malware samples, with AUC of 0.985, and MCC of 0.95; thus, outperforming standard LSTM based methods in these key metrics.

• Asia pacific journal of information systems
• /
• 제20권4호
• /
• pp.25-51
• /
• 2010

This study compares and contrasts the intellectual development of the MIS field in Korea from 2005 to 2009 to that of international trends by using a keyword co-occurrence network analysis of the two flagship journals: APJIS and MIS Quarterly. From 316 research articles in these two journals, 132 unique and most frequently co-occurred keywords were put into analysis. The results of structural equivalence show a mild correlation between APJIS and MIS Quarterly. The e-commerce, trust, and technology adoption are the high frequency keywords in both journals. In Korea e-learning, purchasing, and recommendation systems turn out to be important keywords while outsourcing, research method, quantitative method, design research, information theory, and empirical research are in average international journals. This connotes that the Korean scholarship tends to focus more on practically oriented topics, but the clustering and relational mapping of research topics in each journal show a mild level of overlap with distinctive orientations due to intrinsic disparities depending on the concerned journals' geographical scopes, namely domestic or global.

• Asia Marketing Journal
• /
• 제15권3호
• /
• pp.79-95
• /
• 2013

New product innovation is a process of embodying new knowledge in a product and technology licensing is getting popular as a means to innovations and introduction of new product to the market in today's competitive global market environment. Incumbents often rely on technology licensing to access new product opportunities created by other firms. Prior research has examined various aspects of technology licensing agreements such as specific contract terms of licensing agreements, e.g., distribution of control rights, exclusivity of licensing agreements, cross-licensing, and the scope of licensing agreements. This study aims to provide answers to an important, but under-researched question: why do some incumbents initiate more licensing agreement for exploratory learning while others do it for exploitative learning along the innovation process? We attempt to extend our knowledge of licensing agreements from an organizational learning perspective. Technology licensing as a specific form of interfirm linkages can be initiated with different learning objectives along the process of new product innovation. The exploratory stages of the innovation process such as discovery or research stages involve extensive searches to create new knowledge or capabilities, whereas the exploitative stages of the innovation process such as application or test stages near the commercialization are more focused on developing specific applications or improving their efficiency or reliability. Thus, different stages of the innovation process generate different types of learning and the resulting technological resources. We examine when incumbents as licensees initiate more licensing agreements for exploratory learning objectives and when more for exploitative learning objectives, focusing on two factors that may influence a firm's formation of exploratory and exploitative licensing agreements: 1) its past radical and incremental innovation experience and 2) its internal investments in R&D and marketing. We develop and test our hypotheses regarding the relationship between a firm's radical and incremental new product experience, R&D investment intensity and marketing investment intensity, and the likelihood of engaging in exploratory and exploitive licensing agreements. Using data collected from various secondary sources (Recap database, Compustat database, and FDA website), we analyzed technology licensing agreements initiated in the biotechnology and pharmaceutical industries from 1988 to 2011. The results of this study show that incumbents initiate exploratory rather than exploitative licensing agreements when they have more radical innovation experience and when they invest in R&D activities more intensively; in contrast, they initiate exploitative rather than exploratory licensing agreements when they have more incremental innovation experience and when they invest in marketing activities more intensively. The findings of this study contribute to the licensing and interfirm cooperation studies. First, this study lays a foundation to understand the organizational learning aspect of technology licensing agreements. Second, this study sheds lights on how a firm's internal investments in R&D and marketing are linked to its tendency to initiate licensing agreements along the innovation process. Finally, the findings of this study provide important insight to managers regarding which technologies to gain via licensing agreements. This study suggests that firms need to consider their internal investments in R&D and marketing as well as their past innovation experiences when they initiate licensing agreements along the process of new product innovation.

• 한국정보전자통신기술학회논문지
• /
• 제16권6호
• /
• pp.489-498
• /
• 2023

하드웨어의 성능 및 컴퓨팅 기술의 발전 덕분에 기후환경 변화를 대비하기 위해 기후예측 모델 또한 발전하고 있다. 한국 기상청은 GloSea6를 도입하여 슈퍼컴퓨터를 이용하여 기상 예측을 하고있으며, 각 대학 및 연구 기관에서는 중소규모 서버에서 사용하기 위해 저해상도 결합모델인 Low-GloSea6를 사용하여 기상 연구에 활용하고 있다. 본 논문에서는 중소규모 서버에서의 기상 연구의 원활한 연구를 위해 Low-GloSea6의 Intel VTune Profiler를 사용한 분석을 진행하였으며 1125.987초의 CPU Time을 수행하는 대기모델의 tri_sor_dp_dp 함수를 Hotspot으로 검출하였다. 수치적 연산을 진행하는 기존 함수에 머신러닝 기법의 하나인 비선형 회귀모델을 적용 및 비교하여 머신러닝 적용 가능성을 확인하였다. 기존 tri_sor_dp_dp 함수의 실제 연산되는 값인 1e-3 ~ 1e-20의 범위를 가지는 Output Data인 변수 "Px"를 기준으로 평가하였을때 K-최근접 이웃 회귀 모델은 MAE가 1.3637e-08, SMAPE가 123.2707%로 가장 우수하게 나타났으며 RMSE의 경우 Light Gradient Boosting Machine 회귀 모델이 2.8453e-08로 가장 우수한 성능을 보이는 것으로 측정되었다. 따라서 Low-GloSea6 수행 과정 중 tri_sor_dp_dp 함수의 데이터를 추출 후 비선형 회귀 모델을 적용한 결과로 기존의 tri_sor_dp_dp 함수의 수치적 연산 값과 K-최근접 이웃 회귀 모델을 비교하였을 때 SMAPE가 123.2707%의 오차가 발생하는 것으로 측정되어 기존 모듈의 대체 가능성이 있다는 것을 확인하였다.