• Proceedings of the Korea Information Processing Society Conference
• /
• 2003.11c
• /
• pp.1579-1582
• /
• 2003

본 연구는 컴포넌트 검색을 위해서 컴포넌트를 컨덱스트에 의해 패싯 분류하고, 컨텍스트와 특성들간의 관련값에 대한 통계적 분석에 의해 시소러스를 구축하여 다중 패싯 분류된 컴포넌트를 효율적으로 검색할 수 있는 방법을 제안하였다. 소스 코드로부터 추출된 특성은 카이제곱 방법을 통하여 간소화가 이루어지며, E-SARM 방법을 사용하여 컨텍스트의 자동 검색이 이루어질 수 있도록 하였다. 쿼리에 대해 자동 검색된 컨덱스트에 의해 후보 컴포넌트가 선정되고, 쿼리와 컴포넌트 간의 유사도가 계산됨으로써 컴포넌트가 검색될 수 있도록 하였다. 본 연구는 다중 패싯 분류된 컴포넌트의 검색에 효율적이며, 컴포넌트의 재사용성을 높일 수 있도록 하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11c
• /
• pp.1949-1952
• /
• 2002

본 연구는 객체지향 컴포넌트 검색을 위해서 개선된 Spreading Activation 방법을 이용하여 다중 패싯 분류된 컴포넌트를 효율적으로 검색할 수 있는 방법을 제안하였다. 객체지향 코드 기반의 관계정의를 위해 특성과 컨텍스트 간에 연관관계를 설정하고, 컨덱스트의 자동 추출을 위한 Spreading Activation 방법의 초기 활성값을 정의하였다. 쿼리에 대해 자동 검색된 컨텍스트에 의해 후보컴포넌트가 선정되고, 쿼리와 컴포넌트 간의 신뢰도가 계산됨으로써 컴포넌트가 검색될 수 있도록 하였다. 본 연구는 다중 패싯 분류된 객체지향 컴포넌트의 검색에 효율적이며, 사용자 수작업의 부담을 최대한 감소시켜 컴포넌트의 재사용성을 높일 수 있도록 하였다.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.42 no.1
• /
• pp.47-52
• /
• 2005

The context-aware retrieval method is one of the fundamental characteristics in ubiquitous computing. The essential aims of context-aware retrieval method are retrieving relevant information and delivering information quickly. We propose a new method that retrieves relevant information and delivers information quickly using characteristics of levelized contexts. We extract rules and recommendation information in the near future using context values and rules. Then we prefetch recommendation information in very near future using access score. Our method retrieves relevant information and deliver information quickly by storing only recommendation information to be needed in near future using the characteristics of levelized contexts.

• Journal of Korea Multimedia Society
• /
• v.11 no.1
• /
• pp.61-75
• /
• 2008

A script language, which represents and controls avatar behaviors in a natural language style, is especially remarkable, because it can provide a fast and easy way to develop an animation scenario script. However, the studies that consider avatar behavior interactions with various virtual objects and intuitive interface techniques to design scenario script have been lack. Therefore, we proposed a context-based avatar-object behavior model and layered script language. The model defines context-based elements to solve ambiguity problems that occur in abstract behavior interface and it provides user interface to control avatar in the object-based approach. Also, the proposed avatar behavior script language consisted of a layered structure that represents domain user interface, motion sequence, and implement environment information at each level. Using the proposed methods, the user can conveniently and quickly design an avatar-object behavior scenario script.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.3
• /
• pp.306-310
• /
• 2008

Since the preference for information recommendation service can change according to the context, we should know the user context before providing information recommendation. This paper proposes recommender system that considers multi-user preference in mobile environment and attempted to apply it to restaurant recommendation. To model the preference of individual users in mobile environment, we have used Bayesian network, and restaurant recommendation mostly should consider not an individual user but several users, so this paper has used AHP of multi-criteria decision making process to obtain the preference of several users based on one of individual users. For experiments, we conducted recommendation in 10 different situations, and finally, we confirmed that the proposed system was evaluated as a good one using a usability test of SUS.

• Journal of KIISE:Software and Applications
• /
• v.32 no.7
• /
• pp.650-662
• /
• 2005

A robust repositioning method is required for annotations to always maintain proper positions when original documents were modified. Robust anchoring in the XML document provides better anchoring results when it includes features of structured documents as well as annotated texts. This paper proposes robust annotation anchoring method in XML document. To do this, this work presents annotation information as logical structure trees, and creates candidate anchors by analyzing matching relations between the annotation and document trees. To select the appropriate candidate anchor among many candidate anchors, this work presents several anchoring criteria based on the textual and label context of anchor nodes in the logical structure trees. As a result, robust anchoring is realized even after various modifications of contexts in the structured document.