• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.1
• /
• pp.1-8
• /
• 2007

When confronted with a query, question answering systems endeavor to extract the most exact answers possible by determining the answer type that fits with the key terms used in the query. However, the efficacy of such systems is limited by the fact that the terms used in a query may be in a syntactic form different to that of the same words in a document. In this paper, we present an efficient semantic query expansion methodology based on query patterns in a question category concept list comprised of terms that are semantically close to terms used in a query. The proposed system first constructs a concept list for each question type and then builds the concept list for each question category using a learning algorithm. The results of the present experiments suggest the promise of the proposed method.

• The KIPS Transactions:PartD
• /
• v.10D no.4
• /
• pp.621-638
• /
• 2003

As demand of XML documents in the Web increases, Web service applications that manage XML documents as their resource are increasing. The view mechanism for XML data could be considered for effective in query processing in these Web service applications. If the XML query results are maintained as XML materialized views and their relevant XML query is processed using them, the query response time could be reduced. There are two types of processing an in path expression, which is one of the core features of XML query languages, using XML materialized views. One is the type where the complete query result is obtained from the materialized view, and the other is the type where some of the result is obtained from the materialized view and the rest is from the underlying XML documents. In this paper, we investigate the second type. An XML query in this paper is an XML path expression which is one of the core features of XML query languages. We first describe the storage structures of the XML materialized views derived from the underlying XML documents in the XML repository. Then, we propose the algorithms to decompose a given XML query into the subquery against the materialized view and the subquery against the underlying XML documents, and to integrate the results of these subqueries. Through performance evaluation, we figure out the condition under which our XML query decomposition using materialized views is more effective than the conventional processing.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2000.06a
• /
• pp.190-197
• /
• 2000

본 연구에서는 사용자의 질의에 대해 대답을 제시하는 질의응답시스템의 평가를 위한 테스트컬렉션을 구축하였다. 질의응답시스템 평가를 위한 테스트컬렉션은 207,067개의 문서, 90개의 질의, 각 질의에 대한 적합성 판정 집합으로 구성되어 있다. 문서집합은 신문기사로 SGML형식으로 가공되었고, 질의는 다양한 유형의 질의와 변형질의를 포함한다. 적합성 판정 집합은 각 질의에 대해서 문서에 대답을 포함하는지의 여부에 따라 적합/부적합으로 판정하였고, 적합한 문서에 대해서는 대답을 표시하였다. 본 연구를 통해 구축된 질의응답시스템 평가를 위한 테스트컬렉션은 질의응답시스템의 객관적인 신뢰성 평가를 위한 기반을 마련하였다.

• Journal of the Korean Society for Library and Information Science
• /
• v.47 no.3
• /
• pp.321-342
• /
• 2013

The purpose of this study was to compare and evaluate retrieval effectiveness of three types of image perception using different retrieval methods. Image types included specific, general, and abstract topics. The retrieval method included text only search, query by example (QBE) search, and a hybrid/hybrid search. Thirty-two college students were recruited for searching topics using Google image search system. The search results were compared with One-Way and Two-Way ANOVA. As a result, text search and hybrid search showed advantage when searching for specific and general topics. On the other hand, the QBE search performed better than both the text-only and hybrid search for abstract topics. The results have implications for the implementation of image retrieval systems.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.14 no.2
• /
• pp.159-178
• /
• 2003

Electronic reference service was defined as reference service designed for remote users and identified by a specific link from a library' website. This study examined whether academic libraries deliver web-based electronic reference services to remote users. Particularly, four types of electronic bulletin board services, free bulletin board, FAQ, Q＆A and reference question-response service were measured in terms of 8 types of questions and 12 contents of questions. The result showed that there was few differences among 4 services. Guidelines for effective use of electronic bulletin board services in academic libraries were suggested: standardization of service names, easy access to menu services, definition of each menu, search function for contents and standard web forms.

• Annual Conference on Human and Language Technology
• /
• 2004.10d
• /
• pp.303-307
• /
• 2004

본 논문에서는 서술형 정답을 요구하는 질의에 대해 올바른 서술형 정답을 추출하는 서술형질의응답시스템에 대해 기술한다. 질의응답시스템에서 요구되는 다양한 서술형 정답을 추출하기 위해 정답 유형을 10가지로 정의하였다. 말뭉치로부터 각 서술형 정답 유형에 대한 정답 패턴을 정의하고, 패턴별 제약 규칙 및 각 유형별 패턴적용 순위화 등을 사용하여 정확한 서술형 정답이 추출되도록 하였다. 정답 패턴은 서술형 정답의 구문 구조 및 각 패턴 또는 정답 유형별 실마리 어휘 등으로 구성된다. 현재 학습되지 않은 일반 문서에 대해 59.2%의 서술형 정답 추출 정확도를 보이며, 시스템 성능 향상을 위해 연구가 진행중이다.

• Annual Conference on Human and Language Technology
• /
• 2003.10d
• /
• pp.214-218
• /
• 2003

질의응답 시스템은 인터넷과 같은 실용적 환경에서 사용될 경우, 실제 사용자의 질의는 다양한 유형으로 나타나게 된다. 따라서 실용적인 시스템에서 사용되는 질의는 문장의 형태나 단어의 쓰임에 관계없이 같은 의도를 가진 질의를 같은 유형으로 분류할 수 있는 의문형 문장패턴을 태깅하여 다양한 형태의 자연어로 기술된 문서에서 원하는 응답으로 처리할 수 있는 질의 응답 시스템은 정보 검색 시스템으로서의 가능성을 보여준다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.04a
• /
• pp.816-818
• /
• 2015

본 논문에서는 질의유형을 통한 검색기반 질의응답 시스템을 구현하기 위한 설계방법을 제안한다. 이를 위해 위키피디아 문서의 링크 데이터를 이용하여 색인 대상문서와 데이터베이스를 구축하는 색인 모델과 2-포아송 모델을 이용하여 얻은 문서들을 색인 데이터베이스를 통해 필터링하여 정답 후보문장을 추출하는 검색모델, 키워드 패턴 매칭 기반 질의유형 분류 모델을 설계하였다.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.5
• /
• pp.221-226
• /
• 2024

In information retrieval, queries come in various types, ranging from abstract queries to those containing specific keywords, making it a challenging task to accurately produce results according to user demands. Additionally, search systems must handle queries encompassing various elements such as typos, multilingualism, and codes. Reranking is performed through training suitable documents for queries using DeBERTa, a deep learning model that has shown high performance in recent research. To evaluate the effectiveness of the proposed method, experiments were conducted using the test collection of the Product Search Track at the TREC 2023 international information retrieval evaluation competition. In the comparison of NDCG performance measurements regarding the experimental results, the proposed method showed a 10.48% improvement over BM25, a basic information retrieval model, in terms of search through query error handling, provisional relevance feedback-based product title-based query expansion, and reranking according to query types, achieving a score of 0.7810.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.1
• /
• pp.65-73
• /
• 2008

The telematics data management deals with queries on stream data coming from moving cars. So the stream DBMS should process the large amount of data stream in real-time. In this article, previous research projects are analyzed in the aspects of query processing. And a hybrid model is introduced where query preprocessor is used to process all types of queries in one singe system. Decreasing cost and rapidly increasing Performance of devices may guarantee the utmost parallelism of the hybrid system. As a result, various types of stream DBMS queries could be processed in a uniform and efficient way in a single system.