• Journal of KIISE
• /
• v.43 no.10
• /
• pp.1085-1093
• /
• 2016

Community-based Question Answering system is a system which provides answers for each question from the documents uploaded on web communities. In order to enhance the capacity of question analysis, former methods have developed specific rules suitable for a target region or have applied machine learning to partial processes. However, these methods incur an excessive cost for expanding fields or lead to cases in which system is overfitted for a specific field. This paper proposes a multiple machine learning method which automates the overall process by adapting appropriate machine learning in each procedure for efficient processing of community-based Question Answering system. This system can be divided into question analysis part and answer selection part. The question analysis part consists of the question focus extractor, which analyzes the focused phrases in questions and uses conditional random fields, and the question type classifier, which classifies topics of questions and uses support vector machine. In the answer selection part, the we trains weights that are used by the similarity estimation models through an artificial neural network. Also these are a number of cases in which the results of morphological analysis are not reliable for the data uploaded on web communities. Therefore, we suggest a method that minimizes the impact of morphological analysis by using character features in the stage of question analysis. The proposed system outperforms the former system by showing a Mean Average Precision criteria of 0.765 and R-Precision criteria of 0.872.

• Journal of KIISE:Software and Applications
• /
• v.36 no.4
• /
• pp.306-319
• /
• 2009

A name is a key feature for distinguishing people, but we often fail to discriminate people because an author may have multiple names or multiple authors may share the same name. Such name ambiguity problems affect the performance of document retrieval, web search and database integration. Especially, in bibliography information, a number of errors may be included since there are different authors with the same name or an author name may be misspelled or represented with an abbreviation. For solving these problems, it is necessary to disambiguate the names inputted into the database. In this paper, we propose a method to solve the name ambiguity by using social networks constructed based on the relations between authors. We evaluated the effectiveness of the proposed system based on DBLP data that offer computer science bibliographic information.

• The KIPS Transactions:PartD
• /
• v.9D no.5
• /
• pp.789-800
• /
• 2002

XML is a standard format of web data and is currently used as a prevailing language for exchanging data. Most of the commercial data are stored in a relational database. It is quite important to convert these conventionally stored data into those for exchange and use them in data exchange, or to get the query results effectively by utilizing XQL on XML data which are store in a relational database. Thus, it is absolutely required to have a proper query processing mechanism for XML data and to maintain many XML data properly. Up to now, many cases of researches on the storage and retrieval of XML data have been carried out and under study. But, effective retrieval and storage system for path queries like XQL has yet to be contrived. Thus, in this paper, a schema to store XML data is designed, in which DFS-Numbegering method is used to store data effectively. And an effective path query processing method is also designed and implemented, in which a traditional relational database engine is used. That is, XQL is converted into SQL with a XQL processor if a user makes query XQL in a system. A database system executes SQL, and a XML generator uses a generated record and makes a XML document.

• Journal of Intelligence and Information Systems
• /
• v.16 no.4
• /
• pp.21-41
• /
• 2010

This paper proposes a system that can serve users with appropriate search results through real time filtering, and implemented adaptive user profiling based personalized information retrieval system(PIRS) using users' implicit feedbacks in order to deal with the problem of existing search systems such as Google or MSN that does not satisfy various user' personal search needs. One of the reasons that existing search systems hard to satisfy various user' personal needs is that it is not easy to recognize users' search intentions because of the uncertainty of search intentions. The uncertainty of search intentions means that users may want to different search results using the same query. For example, when a user inputs "java" query, the user may want to be retrieved "java" results as a computer programming language, a coffee of java, or a island of Indonesia. In other words, this uncertainty is due to ambiguity of search queries. Moreover, if the number of the used words for a query is fewer, this uncertainty will be more increased. Real-time filtering for search results returns only those results that belong to user-selected domain for a given query. Although it looks similar to a general directory search, it is different in that the search is executed for all web documents rather than sites, and each document in the search results is classified into the given domain in real time. By applying information filtering using real time directory classifying technology for search results to personalization, the number of delivering results to users is effectively decreased, and the satisfaction for the results is improved. In this paper, a user preference profile has a hierarchical structure, and consists of domains, used queries, and selected documents. Because the hierarchy structure of user preference profile can apply the context when users perfomed search, the structure is able to deal with the uncertainty of user intentions, when search is carried out, the intention may differ according to the context such as time or place for the same query. Furthermore, this structure is able to more effectively track web documents search behaviors of a user for each domain, and timely recognize the changes of user intentions. An IP address of each device was used to identify each user, and the user preference profile is continuously updated based on the observed user behaviors for search results. Also, we measured user satisfaction for search results by observing the user behaviors for the selected search result. Our proposed system automatically recognizes user preferences by using implicit feedbacks from users such as staying time on the selected search result and the exit condition from the page, and dynamically updates their preferences. Whenever search is performed by a user, our system finds the user preference profile for the given IP address, and if the file is not exist then a new user preference profile is created in the server, otherwise the file is updated with the transmitted information. If the file is not exist in the server, the system provides Google' results to users, and the reflection value is increased/decreased whenever user search. We carried out some experiments to evaluate the performance of adaptive user preference profile technique and real time filtering, and the results are satisfactory. According to our experimental results, participants are satisfied with average 4.7 documents in the top 10 search list by using adaptive user preference profile technique with real time filtering, and this result shows that our method outperforms Google's by 23.2%.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.43-61
• /
• 2019

Development of technologies in artificial intelligence has been rapidly increasing with the Fourth Industrial Revolution, and researches related to AI have been actively conducted in a variety of fields such as autonomous vehicles, natural language processing, and robotics. These researches have been focused on solving cognitive problems such as learning and problem solving related to human intelligence from the 1950s. The field of artificial intelligence has achieved more technological advance than ever, due to recent interest in technology and research on various algorithms. The knowledge-based system is a sub-domain of artificial intelligence, and it aims to enable artificial intelligence agents to make decisions by using machine-readable and processible knowledge constructed from complex and informal human knowledge and rules in various fields. A knowledge base is used to optimize information collection, organization, and retrieval, and recently it is used with statistical artificial intelligence such as machine learning. Recently, the purpose of the knowledge base is to express, publish, and share knowledge on the web by describing and connecting web resources such as pages and data. These knowledge bases are used for intelligent processing in various fields of artificial intelligence such as question answering system of the smart speaker. However, building a useful knowledge base is a time-consuming task and still requires a lot of effort of the experts. In recent years, many kinds of research and technologies of knowledge based artificial intelligence use DBpedia that is one of the biggest knowledge base aiming to extract structured content from the various information of Wikipedia. DBpedia contains various information extracted from Wikipedia such as a title, categories, and links, but the most useful knowledge is from infobox of Wikipedia that presents a summary of some unifying aspect created by users. These knowledge are created by the mapping rule between infobox structures and DBpedia ontology schema defined in DBpedia Extraction Framework. In this way, DBpedia can expect high reliability in terms of accuracy of knowledge by using the method of generating knowledge from semi-structured infobox data created by users. However, since only about 50% of all wiki pages contain infobox in Korean Wikipedia, DBpedia has limitations in term of knowledge scalability. This paper proposes a method to extract knowledge from text documents according to the ontology schema using machine learning. In order to demonstrate the appropriateness of this method, we explain a knowledge extraction model according to the DBpedia ontology schema by learning Wikipedia infoboxes. Our knowledge extraction model consists of three steps, document classification as ontology classes, proper sentence classification to extract triples, and value selection and transformation into RDF triple structure. The structure of Wikipedia infobox are defined as infobox templates that provide standardized information across related articles, and DBpedia ontology schema can be mapped these infobox templates. Based on these mapping relations, we classify the input document according to infobox categories which means ontology classes. After determining the classification of the input document, we classify the appropriate sentence according to attributes belonging to the classification. Finally, we extract knowledge from sentences that are classified as appropriate, and we convert knowledge into a form of triples. In order to train models, we generated training data set from Wikipedia dump using a method to add BIO tags to sentences, so we trained about 200 classes and about 2,500 relations for extracting knowledge. Furthermore, we evaluated comparative experiments of CRF and Bi-LSTM-CRF for the knowledge extraction process. Through this proposed process, it is possible to utilize structured knowledge by extracting knowledge according to the ontology schema from text documents. In addition, this methodology can significantly reduce the effort of the experts to construct instances according to the ontology schema.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.40 no.6
• /
• pp.106-117
• /
• 2003

The advanced medical information systems usually consist of loosely-coupled interaction of independent systems, such as HIS/RIS and PACS. To support easier information exchange between these systems and between hospitals, and to support new types of medical service such as teleradiology, it becomes essential to integrate separated medical information and allow them to be exchanged and retrieved through internet. This thesis proposes an integrated medical information system using XML. We analyzed HL7 and DICOM standard formats, and designed an integrated XML DTD. We extracted information from HL7 messages and DICOM files and generated XML document instances and XSL stylesheets based on the proposed XML DTD. We implemented the web interface for the integrated medical information system, which supports data sharing, information exchange and retrieval between two different standard formats. The proposed XML-based integrated medical information system will contribute to solve the problems of current medical information systems, by enabling integration of separated medical informations and by allowing data exchange and sharing through internet. The proposed system with XML is more robust than web-based medical information systems developed by using HTML, because XML itself provides more flexibility and extensibility than HTML.

• The KIPS Transactions:PartB
• /
• v.16B no.3
• /
• pp.225-232
• /
• 2009

Author name disambiguation is essential for improving performance of document indexing, retrieval, and web search. Author name disambiguation resolves the conflict when multiple authors share the same name label. This paper introduces a novel approach which exploits ontologies and WordNet-based category utility for author name disambiguation. Our method utilizes author knowledge in the form of populated ontology that uses various types of properties: titles, abstracts and co-authors of papers and authors' affiliation. Author ontology has been constructed in the artificial intelligence and semantic web areas semi-automatically using OWL API and heuristics. Author name disambiguation determines the correct author from various candidate authors in the populated author ontology. Candidate authors are evaluated using proposed WordNet-based category utility to resolve disambiguation. Category utility is a tradeoff between intra-class similarity and inter-class dissimilarity of author instances, where author instances are described in terms of attribute-value pairs. WordNet-based category utility has been proposed to exploit concept information in WordNet for semantic analysis for disambiguation. Experiments using the WordNet-based category utility increase the number of disambiguation by about 10% compared with that of category utility, and increase the overall amount of accuracy by around 98%.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.4
• /
• pp.412-429
• /
• 2008

As the amount of electronic documents increases rapidly with the growth of the Internet, a parallel search engine capable of handling a large number of documents are becoming ever important. To implement a parallel search engine, we need to partition the inverted index and search through the partitioned index in parallel. There are two methods of partitioning the inverted index: 1) document-identifier based partitioning and 2) keyword-identifier based partitioning. However, each method alone has the following drawbacks. The former is convenient in inserting documents and has high throughput, but has poor performance for top h query processing. The latter has good performance for top-k query processing, but is inconvenient in inserting documents and has low throughput. In this paper, we propose a hybrid partitioning method to compensate for the drawback of each method. We design and implement a parallel search engine that supports the hybrid partitioning method using the Odysseus DBMS tightly coupled with information retrieval capability. We first introduce the architecture of the parallel search engine-Odysseus/parallel-OOSQL. We then show the effectiveness of the proposed system through systematic experiments. The experimental results show that the query processing time of the document-identifier based partitioning method is approximately inversely proportional to the number of blocks in the partition of the inverted index. The results also show that the keyword-identifier based partitioning method has good performance in top-k query processing. The proposed parallel search engine can be optimized for performance by customizing the methods of partitioning the inverted index according to the application environment. The Odysseus/parallel OOSQL parallel search engine is capable of indexing, storing, and querying 100 million web documents per node or tens of billions of web documents for the entire system.