• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2022년도 춘계학술대회
• /
• pp.206-208
• /
• 2022

최근 빅데이터 산업이 큰 폭으로 발전하는 만큼 개인정보 유출로 인한 사생활 침해 문제의 관심도 높아졌다. 자연어 처리 분야에서는 이를 개체명 인식을 통해 자동화하려는 시도들이 있었다. 본 논문에서는 한국어 위키피디아 문서의 본문에서 비식별화 정보를 지닌 문장을 식별해 반자동으로 개체명 인식 데이터를 구축한다. 이는 범용적인 개체명 인식 데이터에 반해 비식별화 대상이 아닌 정보에 대해 학습되는 비용을 줄일 수 있다. 또한, 비식별화 정보를 분류하기 위해 규칙 및 통계 기반의 추가적인 시스템을 최소화할 수 있는 장점을 가진다. 본 논문에서 제안하는 개체명 인식 데이터는 총 12개의 범주로 분류하며 의료 기록, 가족 관계와 같은 비식별화 대상이 되는 정보를 포함한다. 생성된 데이터셋을 이용한 실험에서 KoELECTRA는 0.87796, RoBERTa는 0.88575의 성능을 보였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권9호
• /
• pp.4641-4657
• /
• 2017

Named Data Networking (NDN) has emerged and become one of the most promising architectures for future Internet. However, like traditional IP-based networking paradigm, NDN may not evade some typical network threats such as malicious data aggregates (MDA), which may lead to bandwidth exhaustion, traffic congestion and router overload. This paper firstly analyzes the damage effect of MDA using realistic simulations in large-scale network topology, showing that it is not just theoretical, and then designs a fine-grained MDA mitigation mechanism (MDAM) based on the cooperation between routers via alert messages. Simulations results show that MDAM can significantly reduce the Pending Interest Table overload in involved routers, and bring in normal data-returning rate and data-retrieval delay.

• International Journal of Internet, Broadcasting and Communication
• /
• 제15권1호
• /
• pp.164-170
• /
• 2023

The purpose of this study is to analyze the big data of Adidas brand participating in the Qatar World Cup in 2022 as a FIFA partner to understand useful information, semantic connection and context from unstructured data. Therefore, this study collected big data generated during the World Cup from Adidas participating in sponsorship as a FIFA partner for the 2022 Qatar World Cup and collected data from major portal sites to understand its meaning. According to text mining analysis, 'Adidas' was used the most 3,340 times based on the frequency of keyword appearance, followed by 'World Cup', 'Qatar World Cup', 'Soccer', 'Lionel Messi', 'Qatar', 'FIFA', 'Korea', and 'Uniform'. In addition, the TF-IDF rankings were 'Qatar World Cup', 'Soccer', 'Lionel Messi', 'World Cup', 'Uniform', 'Qatar', 'FIFA', 'Ronaldo', 'Korea', and 'Nike'. As a result of semantic network analysis and CONCOR analysis, four groups were formed. First, Cluster A named it 'Qatar World Cup Sponsor' as words such as 'Adidas', 'Nike', 'Qatar World Cup', 'Sponsor', 'Sponsor Company', 'Marketing', 'Nation', 'Launch', 'Official', 'Commemoration' and 'National Team' were formed into groups. Second, B Cluster named it 'Group stage' as words such as 'Qatar', 'Uruguay', 'FIFA' and 'group stage' were formed into groups. Third, C Cluster named it 'Winning' as words such as 'World Cup Winning', 'Champion', 'France', 'Argentina', 'Lionel Messi', 'Advertising' and 'Photograph' formed a group. Fourth, D Cluster named it 'Official Ball' as words such as 'Official Ball', 'World Cup Official Ball', 'Soccer Ball', 'All Times', 'Al Rihla', 'Public', 'Technology' was formed into groups.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권3호
• /
• pp.1639-1658
• /
• 2019

Named entity recognition (NER) is an important technique for improving the performance of data mining and big data analytics. In previous studies, NER systems have been employed to identify named-entities using statistical methods based on prior information or linguistic features; however, such methods are limited in that they are unable to recognize unregistered or unlearned objects. In this paper, a method is proposed to extract objects, such as technologies, theories, or person names, by analyzing the collocation relationship between certain words that simultaneously appear around specific words in the abstracts of academic journals. The method is executed as follows. First, the data is preprocessed using data cleaning and sentence detection to separate the text into single sentences. Then, part-of-speech (POS) tagging is applied to the individual sentences. After this, the appearance and collocation information of the other POS tags is analyzed, excluding the entity candidates, such as nouns. Finally, an entity recognition model is created based on analyzing and classifying the information in the sentences.

• ETRI Journal
• /
• 제46권1호
• /
• pp.59-70
• /
• 2024

We introduce a high-performance named entity recognition (NER) model for written and spoken language. To overcome challenges related to labeled data scarcity and domain shifts, we use transfer learning to leverage our previously developed KorBERT as the base model. We also adopt a meta-pseudo-label method using a teacher/student framework with labeled and unlabeled data. Our model presents two modifications. First, the student model is updated with an average loss from both human- and pseudo-labeled data. Second, the influence of noisy pseudo-labeled data is mitigated by considering feedback scores and updating the teacher model only when below a threshold (0.0005). We achieve the target NER performance in the spoken language domain and improve that in the written language domain by proposing a straightforward rollback method that reverts to the best model based on scarce human-labeled data. Further improvement is achieved by adjusting the label vector weights in the named entity dictionary.

• 한국산학기술학회논문지
• /
• 제7권2호
• /
• pp.182-187
• /
• 2006

본 논문은 통계 기반 접근 방식인 HMM(Hidden Markov model)과 생물학의 개체명에 관한 온톨로지 정보를 이용한 생물학 문서에서의 개체명(named entity) 경계 인식 방법을 제안한다. 제안하는 방법은 31개의 자질 정보를 이용한 평탄화 기법을 사용하며 생물학 개체명의 계층 정보를 이용하여 HMM의 자료 부족 문제를 완화시킬 수 있도록 하였다. 개체명 경계 인식의 학습과 실험을 위하여 GENIA 코퍼스 ver 2.1을 사용하였으며 개체명 경계 인식 실험을 수행한 결과 모든 부류를 사용한 경우보다 정확도 및 실행 속도가 개선됨을 확인하였다.

• International Journal of Contents
• /
• 제17권4호
• /
• pp.1-15
• /
• 2021

Named entity recognition (NER) is a crucial task for NLP, which aims to extract information from texts. To build NER systems, deep learning (DL) models are learned with dictionary features by mapping each word in the dataset to dictionary features and generating a unique index. However, this technique might generate noisy labels, which pose significant challenges for the NER task. In this paper, we proposed DL-dictionary features, and evaluated them on two datasets, including the OntoNotes 5.0 dataset and our new infectious disease outbreak dataset named GFID. We used (1) a Bidirectional Long Short-Term Memory (BiLSTM) character and (2) pre-trained embedding to concatenate with (3) our proposed features, named the Convolutional Neural Network (CNN), BiLSTM, and self-attention dictionaries, respectively. The combined features (1-3) were fed through BiLSTM - Conditional Random Field (CRF) to predict named entity classes as outputs. We compared these outputs with other predictions of the BiLSTM character, pre-trained embedding, and dictionary features from previous research, which used the exact matching and partial matching dictionary technique. The findings showed that the model employing our dictionary features outperformed other models that used existing dictionary features. We also computed the F1 score with the GFID dataset to apply this technique to extract medical or healthcare information.

• 한국컴퓨터정보학회논문지
• /
• 제24권3호
• /
• pp.113-122
• /
• 2019

In this study, I analyzed the measurement data of the early 20s male who are emphasizing the importance of good clothes in the fashion of body-contact clothes. Through this, I tried to provide basic data necessary for making clothing for early 20s male. Using data from Size Korea's 7th Human Body Survey, 588 people aged 20-25 years were analyzed and classified into four types. Type 1 have a thick and short body, narrow ankle and calf, thin legs. And the hip is not sagged, and height is a little short. So I named it 'short & thick body with bird legs'. Type 2 have a broad shoulder, slim and long body, and no sagging shoulders. So I named it 'slim inverted triangular figure'. Type 3 have a small height, thin and short body, and a thick ankle and calf. So I named it 'short & thin body with thick legs'. Type 4 have a tall height, narrow shoulder, and sagging hip and shoulders. So I named it 'Long triangle'. In order to improve fit of body-contact clothes reflecting the trend of men's wear in recent years, it is necessary to develop clothing prototypes by body type. 20s have the most ideal body shape after completion of growth, but differences in the length, thickness, and thickness of the trunk. This is reflected in the apparel pattern system, and it can be expected to increase consumers' satisfaction if they are used to make excellent ready-to-wear patterns.

• 한국통신학회논문지
• /
• 제40권12호
• /
• pp.2419-2426
• /
• 2015

Named Data Networking (NDN) is one of the powerfully evolving future internet architectures. In this paper installation, configuration and several tests are addressed to show how well and properly our NDN testbed have been prepared and established using NDN platform, in order to have interoperability with global NDN testbed. Global NDN testbed status with our NDN node participation was addressed. To verify one reachability on the NDN connection to global NDN testbed, a latency result is presented using NDN ping test.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권10호
• /
• pp.4054-4071
• /
• 2015

Named-Data Networking (NDN) is one of the promising approaches for the Future Internet to cope with the explosion and current usage pattern of Internet traffic. Content provider mobility in the NDN allows users to receive real-time traffic when the content providers are on the move. However, the current solutions for managing these mobile content providers suffer several issues such as long handover latency, high cost, and non-optimal routing path. In this paper, we survey main approaches for provider mobility in NDN and propose an optimal scheme to support the mobile content providers in the large-scale NDN domain. Our scheme predicts the movement of the provider and uses state information in the NDN forwarding plane to set up an optimal new routing path for mobile providers. By numerical analysis, our approach provides NDN users with better service access delay and lower total handover cost compared with the current solutions.