http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
머신 러닝 기반 시스템을 위한 소프트웨어 엔지니어링 교육
진영택 한국지식정보기술학회 2023 한국지식정보기술학회 논문지 Vol.18 No.3
Modern software systems are increasingly integrating components driven by machine learning technology, harnessing its power to enrich various functionalities such as recommendation engines, anomaly detection, classification tasks, and predictive modeling. These software systems, which incorporate machine learning elements, bear distinctive attributes compared to their conventional counterparts. Regardless, it is of utmost importance that these machine learning-based systems uphold the qualities and standards expected of effective and reliable software. From a software engineering viewpoint, there have been numerous challenges in creating and managing these machine learning-based systems. Many literature reviews, and case analyses have offered certain solutions to address these hurdles. Nonetheless, various literatures and case studies have not been reported for software engineering education that help students adapt to the fast-paced evolution of diverse machine learning development environments. Moreover, traditional software engineering education has often overlooked the unique facets of machine learning development. Consequently, education that encapsulates the synergies and differences between these two arenas is yet to be fully realized. The objective of this paper is to bridge this gap. To achieve this, the study brings forward principles and teaching experiences related to the development of machine learning-based systems as part of the software engineering curriculum.
온라인 제품리뷰의 분석을 위한 오피니언 마이닝 기법의 적용
진영택 한국지식정보기술학회 2021 한국지식정보기술학회 논문지 Vol.16 No.1
Understanding product characteristics and opinions from product reviews provided in news, reviews, and discussions on the web site help individuals make decisions about purchasing products. Companies need to improve product quality and design efficient marketing strategies. For this, opinion mining is required to find and analyze people's opinions from unstructured text reviews. Opinion mining aims to find people's opinions and feelings about certain products and services. It includes a variety of methods and techniques ranging from finding simple opinions on products or services to comparative opinions for products. In particular, comparative opinions provide practical information compared to simple opinions and emotions because people with experiences related to the product or service to be compared mainly provide specific information on the strengths, weaknesses, and differences of the product or service. However, it is not easy to classify comparison opinions from user reviews, and there are many techniques for this. In this study, we present the comparative results of applying various machine learning techniques to specific product reviews for comparative opinion mining. Through this, it is intended to help understand various techniques for classifying comparative opinions and extracting useful information such as the preference of product attributes.
진영택 한국지식정보기술학회 2020 한국지식정보기술학회 논문지 Vol.15 No.6
New technologies such as big data analysis and artificial intelligence are evolving based on convergence with information technology and are affecting all industries. To cultivate problem-solving skills in line with this trend, universities are providing programming education as a basic liberal art course. However, students who take programming courses as liberal arts are having a lot of difficulties in connecting the problem-solving process with actual programming. Even at the stage of putting together basic programming elements and applying them to problem-solving, they tend to feel that they can't do anything or give up. To solve this problem, this paper proposes a method of applying a goal-based scenario to programming education. The goal-based scenario is one of the case-based learning approaches such as problem-based learning and action learning. Based on the scenarios, students can clearly understand the context/situation in which particular knowledge is used and learn to apply the earned knowledge to solving specific problems. Therefore, programming learning based on these scenarios can be used as a tool that can be applied to solving and analyzing similar problems, away from syntax and result-oriented education. In this study, we analyze the existing Python liberal arts programming education and present the effects and considerations of the method of applying the goal-based scenario.
소프트웨어 재사용을 지원하기 위한 재구성 방법론에 관한 연구
진영택 한밭대학교 산업과학기술연구소 1995 논문집 Vol.3 No.-
The development of application program using object oriednted language can be realized by creating objects of an appropriate class for an application domain and constructing the class hierarchy that is representative of that domain by reusing classes in class library. In this paper, to maximize the reusability, an restructuring methodology is presented. The restructuring form on which is mainly focused is the modification of inappropriate inheritance structure and factorization of commonality. Because of static nature of the class, it is difficult to restructure the software without the aid of class design information. And then, the structure of class design information is presented.
진영택(YoungTack Jin),김성렬(SeongRyeol Kim),한환수(HwanSu Han),최종무(JongMu Choi),조유근(YooKun Cho) 한국정보과학회 1993 한국정보과학회 학술발표논문집 Vol.20 No.2
기존의 운영제제에서는 프로세스 개념만을 지원하여 병렬응용의 실행 비용이 과다하다. 따라서 본 논문에서 단일처리기 시스템의 유닉스 운영체제에서 다중스레드 환경을 설계 구현한 연구 결과를 바탕으로 커널의 일부분만을 수정함으로써 큰 부담없이 병렬응용에 대한 지원이 가능함을 확인한다. 기존의 자료구조 (struct proc, struct user)에 대한 수정과 새롭게 정의된 자료구조들(struct task, struct thread)에 대해 살펴 본 후, 커널스레드를 생성 관리하는 과정을 살펴 본다. 또한, 커널스레드와 독립적인 사용자 수준의 스레드의 설계와 구현에 대하여 살펴 본다.
진영택(Young-Taek Jin),임일형(Il-Hyung Lim),최면송(Myeon-song Choi),이승재(Seung-Jae Lee) 대한전기학회 2010 대한전기학회 학술대회 논문집 Vol.2010 No.7
본 논문에서는 6분할 3연계 및 부하균등화를 고려하여 자동화개폐기 위치선정 알고리즘을 제안한다. 현재 배전자동화 시스템은 6분할 3연계 구조를 기준으로 운영하고 있다. 하지만 계통이 복잡하여 운영상의 많은 어려움이 있다. 자동화개폐기 위치선정 알고리즘을 통해 6분할 3연계의 구조와 부하균등화를 이루도록 하는 알고리즘을 제안하였다. 사례연구는 영등포 지점의 3개 변전소 48개 D/L을 대상으로 수행하였다.
Agent 기반 스마트 배전 보호시스템 자동 정정 방법
진영택(Young-Taek Jin),석창주(Chang-Ju Seok),장준형(Joon-Hyung Jang),신동호(Dong-Ho Shin),이승재(Seung-Jae Lee),최면송(Myeon-Song Choi) 대한전기학회 2012 대한전기학회 학술대회 논문집 Vol.2012 No.5
본 논문에서는 스마트 배전을 고려하여 보호기기 및 자동화 개폐기에 Agent 개념을 도입하여 다변화하는 계통구조에 따라 손쉽게 보호협조를 할 수 있고, 고장 구간을 절체 할 수 있는 방안을 제시하고, 보호기기 및 자동화 개폐기에 주소 값을 지정하여 자동으로 정정 할 수 있는 자동 정정 프로그램을 구현하였다.