Young-Hoon Seo

A Korean Part-of-Speech Tagging System Using Resolution Rules for Individual Ambiguous Word

In this paper we present a Korean part-of-speech tagging system using resolution rules for individual ambiguous word. Our system resolves lexical ambiguities by common rules, rules for individual ambiguous word, and statistical approach. We built resolution rules for each word which has several distinct morphological analysis results with a view to enhancing tagging accuracy. Statistical approach based on Hidden Markov Model (HMM) is applied for ambiguous words that are not resolved by the rules. The experiment on the test set shows that the part-of-speech tagging system has high accuracy and broad coverage.

Query generation using semantic features

This paper describes a query generation using semantic features to represent the information demand of users for question answering and information retrieval. One of fundamental reasons why unwanted results are included in responses of all information retrieval systems is because queries do not exactly represent the information demand of users. To solve this problem, a query generaton using the semantic feature is intended to extract semantic features which appear commonly in natural language questions of similar type and utilize them for question answering and information retrieval. We extract semantic features from natural language questions using a grammar and generate queries which represent enough information demands of users using semantic features and syntactic structures. For performance improvement of question answering and information retrieval, we introduce a query-document similarity used to rank documents which include generated queries in the high position. We evaluated our mechanism using 100 queries about a person in the web. There was a notable improvement in the precision at N documents when our approach is applied. Especially, we found that an efficient document retrieval is possible by a question analysis based on semantic features on natural language questions which are comparatively short but fully expressing the information demand of users.

Relevant Keyword Collection using Click-log

The aim of this paper is to collect relevant keywords from clicklog data including user`s keywords and URLs accessed using them. Our main hyphothesis is that two or more different keywords may be relevant if users access same URLs using them. Also, they should have higher relationship when the more same URLs are accessed using them. To validate our idea, we collect relevant keywords from clicklog data which is offered by a portal site. As a result, our experiment shows 89.32% precision when we define answer set to only semantically same words, and 99.03% when we define answer set to broader sense. Our approach has merits that it is independent on language and collects relevant words from real world data.

Korean Part-of-Speech Tagging Using Disambiguation Rules for Ambiguous Word and Statistical information

In this paper we describe a Korean part-of-speech tagging approach using disambiguation rules for ambiguous word and statistical information. Our tagging approach resolves lexical ambiguities by common rules, rules for individual ambiguous word, and statistical approach. Common rules are ones for idioms and phrases of common use including phrases composed of main and auxiliary verbs. We built disambiguation rules for each word which has several distinct morphological analysis results to enhance tagging accuracy. Each rule may have morphemes, morphological tags, and/or word senses of not only an ambiguous word itself but also words around it. Statistical approach based on HMM is then applied for ambiguous words which are not resolved by rules. Experiment shows that the part-of-speech tagging approach has high accuracy and broad coverage.

Question Analysis and Expansion based on Semantics

This paper describes a question analysis and expansion based on semantics for on efficient information retrieval. Results of all information retrieval systems include many non-relevant documents because the index cannot naturally reflect the contents of documents and because queries used in information retrieval systems cannot represent enough information in user`s question. To solve this problem, we analyze user`s question semantically, determine the answer type, and extract semantic features. And then we expand user`s question using them and syntactic structures which are used to represent the answer. Our similarity is to rank documents which include expanded queries in high position. Especially, we found that an efficient document retrieval is possible by a question analysis and expansion based on semantics on natural language questions which are comparatively short but fully expressing the information demand of users.

Concept-Based Question Analysis for an Efficient Document Ranking

This paper describes a concept-based question analysis for an efficient document ranking. Our idea is that we can rank efficiently documents containing answers for questions when we use well-defined concepts because concepts occurred in questions of same answer type are similar. That is, we can retrieve more relevant documents if we know the syntactic and semantic role of each word or phrase in question. For each answer type, we define a concept rule which contains core concepts occurred in questions of that answer type. Concept-based question analysis is a process which tags concepts to morphological analysis result of a users question, determines the answer type, and extracts untagged concepts from it using a matched concept rule. Empirical results show that our concept-based question analysis can rank documents more efficiently than any other conventional approach. Also, concept-based approach has additional merits that it is language universal model, and can be combined with arbitrary conventional approaches.

Concept-based question answering system

This paper describes a concept-based approach for question answering system in which concept rather than keyword makes an important role on question analysis, document retrieval, and answer extraction. Our idea is that we can extract correct answer from various paragraphs with different structures when we use well-defined concepts because concepts occurred in questions of same answer type are similar. We defined a concept rule for each answer type. The concept rule contains core concepts occurred in questions of that answer type. Question analysis module extracts concepts from users question and determines the answer type. Document retrieval module retrieves more relevant documents using extracted concepts. Answer extraction module extracts a probable answer from retrieved documents using concepts. Empirical results show that our concept-based approach can retrieve more relevant documents and extract more accurate answer than any other conventional approach. Also, our approach has additional merits that it is language universal model, and can be combined with arbitrary conventional approaches.

Concept-based Question Analysis for Accurate Answer Extraction

This paper describes a concept-based question analysis to analyze concept which is more important than keyword for the accurate answer extraction. Our idea is that we can extract correct answers from various paragraphs with different structures when we use well-defined concepts because concepts occurred in questions of same answer type are similar. That is, we will analyze the syntactic and semantic role of each word or phrase in a question in order to extract more relevant documents and more accurate answer in them. For each answer type, we define a concept frame which is composed of concepts commonly occurred in that type of questions and analyze user`s question by filling a concept frame with a word or phrase. Empirical results show that our concept-based question analysis can extract more accurate answer than any other conventional approach. Also, concept-based approach has additional merits that it is language universal model, and can be combined with arbitrary conventional approaches.

Query Expansion and Term Weighting Method for Document Filtering

In this paper, we propose a query expansion and weighting method for document filtering to increase precision of the result of Web search engines. Query expansion for document filtering uses ConceptNet, encyclopedia and documents of 10% high similarity. Term weighting method is used for calculation of query-documents similarity. In the first step, we expand an initial query into the first expanded query using ConceptNet and encyclopedia. And then we weight the first expanded query and calculate the first expanded query-documents similarity. Next, we create the second expanded query using documents of top 10% high similarity and calculate the second expanded query- documents similarity. We combine two similarities from the first and the second step. And then we re-rank the documents according to the combined similarities and filter off non-relevant documents with the lower similarity than the threshold. Our experiments showed that our document filtering method results in a notable improvement in the retrieval effectiveness when measured using both precision-recall and F-Measure.