Ad Hoc한 대화 정책을 지원하는 멀티 에이전트 플랫폼에 관한 연구

안형준,Ahn, Hyung-Jun 한국정보처리학회 2004 정보처리학회논문지D Vol.11 No.5

Multi-agent systems have been developed for supporting intelligent collaboration of distributed and independent software entities and are be-ing widely used for various applications. For the collaboration among agents, conversation policies (or interaction protocols) mutually agreed by agents are used. In today's dynamic electronic market environment, there can be frequent changes in conversation policies induced by the changes in transaction methods in the market, and thus, the importance of ad hoc conversation policies is increasing. In existing agent platforms, they allow the use of only several standard or fixed conversation policies, which requires inevitable re implementation for ad hoc conversation policies and leads to inefficiency and intricacy. This paper designs an agent platform that supports ad hoc conversation policies and presents the prototype implementation. The suggested system includes an exchangeable and interpretable conversation policy model, a meta conversation procedure for exchanging new conversation policies, and a mechanism for performing actual transactions with exchanged conversation policies in run time in an adaptive way. 멀티 에이전트 시스템은 분산 환경에서 독립된 소프트웨어 개체들의 지능형 협력 작업을 지원하기 위해 개발되었으며 많은 응용 환경에서 사용되어지고 있다. 멀티 에이전트 시스템의 협력을 위해서는 에이전트간에 미리 협의된 형태의 프로토콜인 대화 정책(Conversation Policy; Interaction Protocol)이 사용된다. 현재의 동적인 인터넷 전자시장 환경에서는 거래의 형태가 변화함에 따라 대화 정책이 자주 바뀔 수 있으며 따라서 Ad Hoc한 형태의 대화 정책의 중요성이 커지고 있다. 기존의 에이전트 플랫폼은 몇 개의 표준 혹은 미리 정해진 대화 정책만을 허용함으로써 Ad Hoc한 대화 정책에 대해서는 에이전트 시스템을 일부 새로 구현해주어야 하는 번거로움과 비효율성이 존재한다. 본 연구에서는 그러한 Ad Hoc한 대화 정책을 지원하는 에이전트 플랫폼을 설계하며 그 프로토타입 시스템을 제시한다. 제시된 시스템은 교환 및 해석을 위한 대화 정책 모델, 새로운 대화 정책을 처리하기 위한 메타 대화 처리 부분, 그리고 대화 정책을 런타임시에 교환하고 해석하여 실제 에이전트들이 적응성 있는 상거래 및 협력을 할 수 있도록 하는 메커니즘을 포함한다.

국내 건설회사의 지속가능성장률 분석을 통한 건설 산업 특성에 관한 연구 - Higgins모형과 Babcock모형을 중심으로 -

안형준,최청균,김주형,김재준,Ahn, Hyoung-Jun,Choi, Chung-Kyun,Kim, Ju-Hyung,Kim, Jae-Jun 대한건축학회 2011 대한건축학회논문집 Vol.27 No.5

The Ministry of Construction and Transportation announced the Earned Value Management Since the global financial crisis caused by US sub-prime mortgage turmoil in Sep. 2008, construction industry is under serious difficulties in the management at present about two years to the financial crisis, which are matured PF(Project Financing) loans and at least 50 trillion won of locked up funds by unsold houses, due to the industrial characteristic declining more rapidly than other industries in recession. For this difficulty shown in construction industry, it established the causes not only the depression in construction market overtaken followed by the global financial crisis but also the business strategy of local construction companies excessively weighted towards external growth. In today's rapidly changing society, such business strategy overly focused on growth only may threaten even the corporate sustainability. Therefore, local construction companies must meet the changes in customer demands and the market and they must focus on the establishment of efficient strategies through the corporate strategic management at the same time. Based on this background, this study analyzed the sustainable growth rate with accounting information of 70 construction companies listed to the market of Korea Stock Exchange during the period from 2005 to 2009 and then analyzed the characteristics shown in those companies classified into the good company group and the bad company group for the corporate activity of construction companies by which the reasonable growth can be maintained while the growth-oriented strategy is still valid.

방사선 투시를 이용한 개폐구 운동에 따른 수직 고경 계측점의 안정성에 관한 연구

안형준,김창회,김영수,김용호,Ahn, Hyung-Jun,Kim, Chang-Whe,Kim, Yung-Soo,Kim, Yong-Ho 대한치과보철학회 1999 대한치과보철학회지 Vol.37 No.4

The current clinical technique for occlusal vertical dimension recording is based on marking the skin reference points on the patient's face and measuring between these points using caliper-like device. And it is difficult to achieve reliable measurements by this technique because of movable soft tissue. The purpose of this study is to reveal the stability of skin reference points by comparing the relative movement between extra-oral skin reference points and intra-oral reference points using X-ray fluoroscope. 10 test subjects were divided into 2 groups : Group I (natural dentition) and Group II (denture-wearer whose vertical dimension was lost) and Group III consists of identical test subjects to Group II with their upper denture removed and record base inserted. Attaching the 3 mm diameter steel ball to nose tip, lower lip, chin and to existing denture (or record base), fluoroscopic examination and recording were taken during 2 jaw opening and closing movements. After subsequent digitization using personal computer, 1219 still pictures with 0.1 second interval were made. Using the 2 dimensional graphic software, measurements between reference points were executed. Dividing the entire jaw movement into 3 ranges (total, 1st half opening, 2nd half opening), rate of movement and relative movement between extra-oral and intra-oral reference points were calculated and statistically analyzed. The results of this study are as follows. 1 Within the same experimental group, no statistical difference was found in the stability of skin reference between lower lip point and chin point during total range of jaw opening and closing movement (p>.05) 2. In the first half range of jaw opening, statistical difference was found between Group I (natural dentition) and Group II (denture wearer) (p<.05) Group I has greater skin reference stability than Group II. 3. In the first half range of jaw opening, statistical difference was found between Group I and Group III (record base wearer) (p<.05). Group I has greater skin reference stability than Group III. 4. In the first half range of jaw opening, no statistical difference was found in the stability of skin reference between Group II and Group III (p>.05). 5. In the second half range of jaw opening, no statistical difference was found in the stability of skin reference between any experimental groups (p>.05). 6. In patients with their occlusal vertical dimension lost, employing other measuring references rather than skin is recommended because of low stability.

인터넷 상점에서의 내용기반 추천을 위한 상품 및 고객의 자질 추출 성능 비교

안형준,김종우,Ahn Hyung-Jun,Kim Jong-Woo 한국정보처리학회 2006 정보처리학회논문지D Vol.13 No.2

인터넷 쇼핑몰에서의 상품 추천을 위해 널리 사용되는 방식 중 한 가지는 상품의 특성과 고객의 특성을 비교하여 고객에 맞는 상품을 추천하는 방식이다. 이 방식은 상품이나 고객의 특성을 표현하는 자질(Feature)의 개수가 많을수록 그 중에 어떤 자질을 선택해야 더 좋은 추천 성과를 가져올 수 있는지 파악해 내는 것이 추천의 효과 및 효율성 측면에서 중요하지만 아직까지 충분히 연구되지 않은 실정이다. 본 연구에서는 인터넷 서점에서의 가상 구매실험을 바탕으로 사용자가 구매한 책 들에서 사용자를 잘 나타낼 수 있는 자질을 선택하는 방식에 대해서 벡터 스페이스 모형, TFIDF(Term Frequency-Inverse Document Frequency), Mutual Information, SVD(Singular Value Decomposition) 방식 등을 활용하여 실험하고 그 결과를 비교해본다. 실험 결과 SVD를 응용한 자질 추출 기법이 가장 좋은 성능을 나타내었다. One of the widely used methods for product recommendation in Internet storefronts is matching product features against target customer profiles. When using this method, it's very important to choose a suitable subset of features for recommendation efficiency and performance, which, however, has not been rigorously researched so far. In this paper, we utilize a dataset collected from a virtual shopping experiment in a Korean Internet book shopping mall to compare several popular methods from other disciplines for selecting features for product recommendation: the vector-space model, TFIDF(Term Frequency-Inverse Document Frequency), the mutual information method, and the singular value decomposition(SVD). The application of SVD showed the best performance in the analysis results.

검색 정확도와 상품 추천 효과의 상호 관계

안형준(Hyung Jun Ahn) 한국정보기술학회 2012 한국정보기술학회논문지 Vol.10 No.10

Most studies so far about recommendation systems have used empirical approaches using data sets or experiments. Different from the studies, this paper presents a theoretical model of online shopping and recommendation systems, with which the effectiveness of recommendation systems is analyzed. The proposed model assumes that consumers pursue efficiency of decision making, and is used to test whether the effectiveness of recommendation systems is affected by the characteristic of Internet stores, the accuracy of the search system. The analysis shows that recommendation system is not always effective, and that it becomes relatively more effective as the accuracy of search gets lower.

SNS의 이벤트와 텍스트의 언어심리학적 특성 간의 관계

안형준(Hyung Jun Ahn) 한국정보기술학회 2020 한국정보기술학회논문지 Vol.18 No.5

With the increasing use of SNS, analyzing texts on SNS is an important means of understanding users. This study assumed that psycho-linguistic characteristics of texts can have a significant relationship with the events on SNS, where an event is an important topic of interest that is mentioned or shared by many users every day. In order to analyze the relationship, this study used the theories in psychology such as construal level, self-construal, and regulatory focus, together with a psycho-linguistic text analysis tool. Specifically, the psychological characteristics of user texts were analyzed, which then were evaluated for possible impact on user behavior such as mentions, using hash tags, and sharing URLs. It was also analyzed whether the characteristics are related to the topics of events. The results of binary logit and deep learning analyses respectively showed significant results.

네트워크 분석을 통한 브랜드 이미지 텍스트의 키워드 추출

안형준(Hyung-Jun Ahn) 한국정보기술학회 2012 한국정보기술학회논문지 Vol.10 No.2

Consumers exchange various opinions online using unstructured texts about brands, products, and services. The texts provide very useful information to companies for analyzing consumers reaction about products and brands. In line with this, this study conducted interviews to collect texts about the perceived image of a global communication device manufacturer, and performed network analysis of the texts to extract keywords. The keywords were compared with those directly provided by the interviewees to test the relevancy. The results showed that using network analysis with the Eigenvector centrality provides more relevant keywords than the simple frequency-based method, and that the performance varies depending on how the word-networks are constructed.

한국인 트위터 사용자들의 트위터 사용 행태와 성격 간의 관계

안형준(Hyung Jun Ahn) 한국정보기술학회 2016 한국정보기술학회논문지 Vol.14 No.1

Although abundant research has been conducted so far about social network services with the growing importance of them, domestic studies about Korean users are relatively limited. Especially, it is hard to find any study that investigates the relationship between the usage behavior of social network services and the personality of the users. This study analyzed the relationship between the behavior of Korean Twitter users and their personality. For the study, a survey was conducted for university students to collect their personality information. Their usage data were also collected using the open API(Application Programming Interface) provided by Twitter. The data consist of in total 34 variables that are either linguistic or non-linguistic in their nature. Factor analysis was conducted for the variables and 10 factors were identified. A correlation analysis was then performed between the factors and the personality variables. It was found that two of the factors were significantly related with extraversion, and two with openness.

경험의 유형과 블로그 글의 특성 간의 관계 분석

안형준(Hyung Jun Ahn),하영목(Youngmok Ha) 한국정보기술학회 2017 한국정보기술학회논문지 Vol.15 No.2

With the growing importance of social media, it is imperative to analyze it to understand the users. Users share useful information and their experience through social media, where much of what is shared is in the form of texts. This study focused on blogs and aimed to test whether the psycho-linguistic characteristics of blog texts vary with the subject or the type of experience of the texts. For this goal, blog texts about four different types of experience, Go, skiing, reading, and musical were collected through the search API of the Tistory blog service. The analysis of the texts showed that various psycholinguistic characteristics of the texts are different across the four categories of the texts. Moreover, the machine learning experiment using the characteristics for automatic text classification showed significant performance. Specifically, the ensemble method, based on functional tree and bagging appeared to be most effective in classification.

데이터 마이닝 및 통계 기법을 활용한 색상과 키워드 간의 관계 분석

안형준(Hyung Jun Ahn) 한국정보기술학회 2013 한국정보기술학회논문지 Vol.11 No.12

This study proposes a method of analyzing data that combine images and texts, which can often be found online such as in social network services, so that the relationship between colors of the images and the key words of the texts can be identified. For this goal, dominant colors of the images and the key words extracted from the texts are clustered respectively. The association between the clusters are then identified with cross-tabulation, heat map analysis, and correspondence analysis. Sample data are collected to test the usefulness of the approach. The results show clear relationship between the key words and the colors.