• Smart Media Journal
• /
• v.11 no.3
• /
• pp.31-38
• /
• 2022

In the document extraction summary study, various methods for selecting important sentences based on the relationship between sentences were proposed. In the Korean document summary using the summation similarity of sentences, the summation similarity of the sentences was regarded as the amount of sentence information, and the summary sentences were extracted by selecting important sentences based on this. However, the problem is that it does not take into account the various importance that each sentence contributes to the entire document. Therefore, in this study, we propose a document extraction summary method that provides a summary by selecting important sentences based on the amount of quantitative and semantic information in the sentence. As a result, the extracted sentence agreement was 58.56% and the ROUGE-L score was 34, which was superior to the method using only the combined similarity. Compared to the deep learning-based method, the extraction method is lighter, but the performance is similar. Through this, it was confirmed that the method of compressing information based on semantic similarity between sentences is an important approach in document extraction summary. In addition, based on the quickly extracted summary, the document generation summary step can be effectively performed.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.22 no.5
• /
• pp.656-662
• /
• 2012

In this paper, we describe a new summarizing method based on a graph-based and a sense-based analysis. In the graph-based analysis, we convert sentences in a document into word vectors and calculate the similarity between each sentence using LSA. We reflect this similarity of sentences and the rarity scores of words in sentences to define weights of edges in the graph. Meanwhile, in the sense-based analysis, in order to determine the sense of words, subjectivity or objectivity, we built a database which is extended from the golden standards using Wordnet. We calculate the subjectivity of sentences from the sense of words, and select more subjective sentences. Lastly, we combine the results of these two methods. We evaluate the performance of the proposed method using classification games, which are usually used to measure the performances of summarization methods. We compare our method with the MS-Word auto-summarization, and verify the effectiveness of ours.

• Smart Media Journal
• /
• v.10 no.1
• /
• pp.39-46
• /
• 2021

This study is about a method of extracting a summary from a news article in consideration of the importance of each sentence constituting the article. We propose a method of calculating sentence importance by extracting the probabilities of topic sentence, similarity with article title and other sentences, and sentence position as characteristics that affect sentence importance. At this time, a hypothesis is established that the Topic Sentence will have a characteristic distinct from the general sentence, and a deep learning-based classification model is trained to obtain a topic sentence probability value for the input sentence. Also, using the pre-learned ELMo language model, the similarity between sentences is calculated based on the sentence vector value reflecting the context information and extracted as sentence characteristics. The topic sentence classification performance of the LSTM and BERT models was 93% accurate, 96.22% recall, and 89.5% precision, resulting in high analysis results. As a result of calculating the importance of each sentence by combining the extracted sentence characteristics, it was confirmed that the performance of extracting the topic sentence was improved by about 10% compared to the existing TextRank algorithm.

• The Journal of the Korea Contents Association
• /
• v.12 no.9
• /
• pp.9-15
• /
• 2012

A conventional event sentence extraction research doesn't learn the 3W features in the learning step and applies the rule on whether the 3W feature exists in the extraction step. This paper presents a sentence weight based event sentence extraction method that calculates the weight of the 3W features in the learning step and applies the weight of the 3W features in the extraction step. In the experimental result, we show that top 30% features by the $TF{\times}IDF$ weighting method is good in the feature filtering. In the real estate domain of the public issue, the performance of sentence weight based event sentence extraction method is improved by who and when of 3W features. Moreover, In the real estate domain of the public issue, the sentence weight based event sentence extraction method is better than the other machine learning based extraction method.

• Annual Conference on Human and Language Technology
• /
• 2009.10a
• /
• pp.31-34
• /
• 2009

본 연구의 목표는 비교 문장들을 일곱 가지 유형으로 자동 분류하는 것으로서, 비교 문장 추출, 비교 문장 유형 분류, 유형별 비교 관계 분석으로 이어지는 비교마이닝 세 단계 중 두 번째 과제이다. 본 연구에서는 변환 기반 학습(Transformation-based Learning) 기법을 이용한다. 자연어 처리 분야 여러 부문에서 사용되고 있는 변환 기반 학습은 오류를 감소시키는 최적의 규칙을 자동으로 생성하여 정답을 찾는 규칙 기반 학습 방법이다. 웹상의 다양한 도메인에서 추출한 비교 문장들을 대상으로 실험한 결과, 일곱 가지 비교 문장 유형을 분류하는데 있어서 정확도 80.01%의 우수한 성능을 산출하였다.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.50-54
• /
• 2018

관계 추출을 위한 원격 지도 학습은 사람의 개입 없이 대규모 데이터를 생성할 수 있는 효율적인 방법이다. 그러나 원격 지도 학습은 노이즈 데이터 문제가 있으며, 노이즈 데이터는 두 가지 유형으로 나눌 수 있다. 첫 번째는 관계 표현 자체가 없는 문장이 연결된 경우이고, 두 번째는 관계 표현은 있는 문장이지만 다른 관계 표현도 함께 가지는 경우이다. 주로 문장의 길이가 길고 복잡한 문장에서 두 번째 노이즈 데이터 유형이 자주 발견된다. 본 연구는 두 번째 경우의 노이즈를 줄임으로써 관계 추출 모델의 성능을 향상시키기 위해 확장된 최단 의존 경로를 사용하는 CNN 기반 관계 추출 모델을 제안한다. 본 논문에서 제안한 방법의 우수성을 입증하기 위해, 한국어 위키피디아와 DBpedia 기반의 원격 지도 학습 데이터를 수집하여 평가한 결과, 본 논문에서 제안한 방법이 위 문제를 해결하는데 효과적이라는 것을 확인하였다.

• Annual Conference on Human and Language Technology
• /
• 1995.10a
• /
• pp.3-7
• /
• 1995

한국어에서의 명사구 색인을 위한 기존의 방법들은 주로 간단한 규칙을 이용하여 왔고 그 결과 문장에 존재하는 모든 명사구를 추출하지 못했다. 이를 해결하기 위하여 본 논문에서는 개념 기반 명사구 색인 방법을 제안한다. 하나의 문장은 하나 이상의 개념으로 이루어져 있으므로, 명사구 추출은 개념을 고려하여 이루어져야 바람직하다 문장은 구문적으로 하나 이상의 내포문으로 이루어져 있다. 일반적으로 내포문 단위 내의 용어들이 나타내는 각각의 개념들은 서로 높은 연관성을 가진다. 그러므로 문장이 가지는 개념의 상이성을 내포문의 개념 상이성으로 축소할 수 있다. 문장을 내포문 단위로 분할하기 위하여 의존 문법을 기반한 구문분석과 공기정보를 이용한다. 특히 공기정보는 원거리 의존관계(long distance dependency)를 결정하여 한 내포문에 속함을 밝혀내는 데 도움을 준다. 이러한 내포문 내의 의존관계를 이용하여 명사구를 추출한다.

• Journal of the Korean Society for information Management
• /
• v.26 no.1
• /
• pp.305-320
• /
• 2009

The news pages provided through the web contain unnecessary information. This causes low performance and inefficiency of the news processing system. In this study, news content extraction methods, which are based on sentence identification and block-level tags news web pages, was suggested. To obtain optimal performance, combinations of these methods were applied. The results showed good performance when using an extraction method which applied the sentence identification and eliminated hyperlink text from web pages. Moreover, this method showed better results when combined with the extraction method which used block-level. Extraction methods, which used sentence identification, were effective for raising the extraction recall ratio.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.176-178
• /
• 2018

트위터는 최대 140자의 단문을 주고받는 소셜 네트워크 서비스이다. 트위터의 해시 태그는 주로 문장의 핵심 단어나 주요 토픽 등을 링크하게 되는데 본 논문에서는 이러한 정보를 이용하여 키워드 추출에 활용한다. 문장을 Character CNN, Bi-LSTM을 통해 문장 표현을 얻어내고 각 Span에서 이러한 문장 표현을 활용하여 Span 표현을 생성한다. Span 표현을 이용하여 각 Span에 대한 Score를 얻고 높은 점수의 Span을 이용하여 키워드를 추출한다.

• Smart Media Journal
• /
• v.6 no.1
• /
• pp.22-29
• /
• 2017

Due to the rapid advancement and distribution of smart devices of late, document data on the Internet is on the sharp increase. The increment of information on the Web including a massive amount of documents makes it increasingly difficult for users to understand corresponding data. In order to efficiently summarize documents in the field of automated summary programs, various researches are under way. This study uses TextRank algorithm to efficiently summarize documents. TextRank algorithm expresses sentences or keywords in the form of a graph and understands the importance of sentences by using its vertices and edges to understand semantic relations between vocabulary and sentence. It extracts high-ranking keywords and based on keywords, it extracts important sentences. To extract important sentences, the algorithm first groups vocabulary. Grouping vocabulary is done using a scale of specific weight. The program sorts out sentences with higher scores on the weight scale, and based on selected sentences, it extracts important sentences to summarize the document. This study proved that this process confirmed an improved performance than summary methods shown in previous researches and that the algorithm can more efficiently summarize documents.