• Journal of KIISE
• /
• v.41 no.12
• /
• pp.1058-1065
• /
• 2014

In order to assist user's who are in the process of executing a search, a query expansion method suggests keywords that are related to an input query. Recently, several studies have suggested keywords that are identified by finding domains using a clustering method over the documents that are retrieved. However, the clustering method is not relevant when presenting various domains because the number of clusters should be fixed. This paper proposes a method that suggests keywords by finding various domains related to the input queries by using a community detection algorithm. The proposed method extracts words from the top-30 documents of those that are retrieved and builds communities according to the word graph. Then, keywords representing each community are derived, and the represented keywords are used for the query expansion method. In order to evaluate the proposed method, we compared our results to those of two baseline searches performed by the Google search engine and keyword recommendation using TF-IDF in the search results. The results of the evaluation indicate that the proposed method outperforms the baseline with respect to diversity.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.5
• /
• pp.1096-1101
• /
• 2005

An efficient method of estimating the spatial gaps between the connected components has been prposed to separatethe individual words from a handwritten legal amount on bank check. Owing to the inherent problem of underestimation or overestimation, the previous gap measures have much difficulty in being applied to the legal amounts that usually include the great shape variability by writer's unconstrained writing style and touching or irregular gaps between words by space limitation. In order to alleviate such burden and improve word separation performance, we have developed a modified version of each distance measure. Through a series of word separation experiments, we found that the modified distance measures show a better performance with over $2-3\%$ of the word separation rate than their corresponding original distance measures.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.7
• /
• pp.319-324
• /
• 2006

We Propose two new methods for non-keyword modeling to improve the performance of speaker- and vocabulary-independent keyword spotting system. The first method is decision tree clustering of monophone at the state level instead of monophone clustering method based on K-means algorithm. The second method is multi-state multiple mixture modeling at the syllable level rather than single state multiple mixture model for the non-keyword. To evaluate our method, we used the ETRI speech DB for training and keyword spotting test (closed test) . We also conduct an open test to spot 100 keywords with 400 sentences uttered by 4 speakers in an of fce environment. The experimental results showed that the decision tree-based state clustering method improve 28%/29% (closed/open test) than the monophone clustering method based K-means algorithm in keyword spotting. And multi-state non-keyword modeling at the syllable level improve 22%/2% (closed/open test) than single state model for the non-keyword. These results show that two proposed methods achieve the improvement of keyword spotting performance.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.12
• /
• pp.477-484
• /
• 2018

Due to the advancement of computer and information technologies, numerous papers have been published. As new research fields continue to be created, users have a lot of trouble finding and categorizing their interesting papers. In order to alleviate users' this difficulty, this paper presents a method of grouping similar papers and clustering them. The presented method extracts primary keywords from the abstracts of each paper by using TF-IDF. Based on TF-IDF values extracted using K-means clustering algorithm, our method clusters papers to the ones that have similar contents. To demonstrate the practicality of the proposed method, we use paper data in FGCS journal as actual data. Based on these data, we derive the number of clusters using Elbow scheme and show clustering performance using Silhouette scheme.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.10
• /
• pp.731-738
• /
• 2013

We analyze public opinion via a sentiment analysis of tweets collected by using recent topic keywords extracted from newspaper articles. Newspaper articles collected within a certain period of time are clustered by using K-means algorithm and topic keywords for each cluster are extracted by using term frequency. A sentiment analyzer learned by a machine learning method can classify tweets according to their polarity values. We have an assumption that tweets collected by using these topic keywords deal with the same topics as the newspaper articles mentioned if the tweets and the newspapers are generated around the same time. and we tried to verify the validity of this assumption.

• The KIPS Transactions:PartB
• /
• v.13B no.6 s.109
• /
• pp.591-600
• /
• 2006

This paper describes the development and implementation of a algorithm to decompose word images from image regions mixed text/graphics in document images using statistical analyses. To decompose word images from image regions, the character components need to be separated from graphic components. For this process, we propose a method to separate them with an analysis of box-plot using a statistics of structural components. An accuracy of this method is not sensitive to the changes of images because the criterion of separation is defined by the statistics of components. And then the character regions are determined by analyzing a local crowdedness of the separated character components. finally, we devide the character regions into text lines and word images using projection profile analysis, gap clustering, special symbol detection, etc. The proposed system could reduce the influence resulted from the changes of images because it uses the criterion based on the statistics of image regions. Also, we made an experiment with the proposed method in document image processing system for keyword spotting and showed the necessity of studying for the proposed method.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06c
• /
• pp.319-322
• /
• 2011

시각 피처를 사용한 이미지 표현은 이미지 검색 분야에서 이미 광범위하게 사용되고 있다. 특히 이미지 자체에 태깅이 되어있지 않거나 다른 추가 정보가 없는 경우에는 이미지 콘텐츠자체의 정보만으로 검색하기 위해서는 이러한 전처리가 필수적이다. 이미지로 부터 얻어진 시각적 피처들이 시각 단어로 사용되기 위해서는 k-means 와 같은 군집 알고리즘을 통한 시각적 피처의 양자화를 위한 전처리가 필요한데, 시각 단어의 개수 k를 정하는데 모호함이 있다. 본 논문에서는 임의의 k를 사용하더라도, 대표적 토픽 모델링 기법인 LDA (Latent Dirichlet Allocation)를 사용하여 데이터의 차원을 줄이게 되면 여러개의 시각적 단어들의 조합을 각각의 토픽이 나타낼 수 있게 됨을 이미지 검색 성능으로써 확인해 보고, 이러한 방법을 사용하면 표현형의 사이즈를 줄일 수 있고, 검색에 있어서도 이미지의 유사성을 더욱 효과적으로 표현할 수 있음을 확인해 본다.

• Journal of Information Management
• /
• v.42 no.4
• /
• pp.137-154
• /
• 2011

A headword finding system is defined as an information retrieval system using a word gloss as a query. We use the gloss as a document in order to implement such a system. Generally the gloss is very short in length and then makes very difficult to find the most proper headword for a given query. To alleviate this problem, we expand the document using the concept of query expansion in information retrieval. In this paper, we use 2 document expansion methods : gloss expansion and similar word expansion. The former is the process of inserting glosses of words, which include in the document, into a seed document. The latter is also the process of inserting similar words into a seed document. We use a featureless clustering algorithm for getting the similar words. The performance (r-inclusion rate) amounts to almost 100% when the queries are word glosses and r is 16, and to 66.9% when the queries are written in person by users. Through several experiments, we have observed that the document expansions are very useful for the headword finding system. In the future, new measures including the r-inclusion rate of our proposed measure are required for performance evaluation of headword finding systems and new evaluation sets are also needed for objective assessment.

• Journal of Korea Entertainment Industry Association
• /
• v.13 no.4
• /
• pp.173-184
• /
• 2019

The data processing of this study is based on the word data search in social media through textom and the big data analysis is carried out and three areas (2019 Gwangju World Swimming Championships, 2019 Gwangju World Swimming Masters Competition, 2019 World Swimming Championships Problem) was consistently handled through data collection and refinement in the web environment. We applied the collected words to the program of Ucinet6, visualized them, and conducted a CONCOR analysis to grasp the similar relationship of words and to identify the cluster of common factors. As a result of the analysis, the clusters related to the 2019 Gwangju World Swimming Championships mainly consisted of four major areas of recognition and perception, mainly searching for operational aspects related to the swimming championship, and the community related to the 2019 Gwangju World Swimming Masters Competition Is mainly searched for the promotion of the Masters Competition and the aspect of the competition divided into two areas of major recognition and peripheral recognition. The cluster related to the problems of the 2019 Gwangju World Swimming Championships is divided into five areas, And they are mainly searching for the place, operation, institution, event, etc. of the problem of the swimming championship.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.264-269
• /
• 2021

상호참조해결은 주어진 문서에서 상호참조해결의 대상이 될 수 있는 멘션을 추출하고, 같은 개체를 의미하는 멘션 쌍 또는 집합을 찾는 자연어처리 작업이다. 하나의 멘션 내에 멘션이 될 수 있는 다른 단어를 포함하는 중첩 멘션은 순차적 레이블링으로 해결할 수 없는 문제가 있다. 본 논문에서는 이러한 문제를 해결하기 위해 멘션의 시작 단어의 위치를 여는 괄호('('), 마지막 위치를 닫는 괄호(')')로 태깅하고 이 괄호들을 예측하는 멘션 탐지 모델과 멘션 탐지 모델에서 예측된 멘션을 바탕으로 포인터 네트워크를 이용하여 같은 개체를 나타내는 멘션을 군집화하는 상호참조해결 모델을 제안한다. 실험 결과, 4개의 영어 대화 데이터셋에서 멘션 탐지 모델은 F1-score (Light) 94.17%, (AMI) 90.86%, (Persuasion) 92.93%, (Switchboard) 91.04%의 성능을 보이고, 상호참조해결 모델에서는 CoNLL F1 (Light) 69.1%, (AMI) 57.6%, (Persuasion) 71.0%, (Switchboard) 65.7%의 성능을 보인다.