• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.1
• /
• pp.160-173
• /
• 2021

Recently, research on the development of autonomous vehicles has increased worldwide. Moreover, a means to identify and analyze the characteristics of traffic accidents of autonomous vehicles is needed. Accordingly, traffic accident data of autonomous vehicles are being collected in California, USA. This research examined the characteristics of traffic accidents of autonomous vehicles. Primarily, traffic accident data for autonomous vehicles were analyzed, and the text data used text-embedding techniques to derive major keywords and four topics. The methodology of this study is expected to be used in the analysis of traffic accidents in autonomous vehicles.

• Journal of Korean Society for Quality Management
• /
• v.50 no.4
• /
• pp.717-734
• /
• 2022

Purpose: The purpose of this study is to effectively detect violations that occur simultaneously against Good Manufacturing Practice, which were concealed by drug manufacturers. Methods: In this study, we present an analysis framework for analyzing regulatory violation patterns using Association Rule Mining (ARM), Text Mining, and t-distributed Stochastic Neighbor Embedding (t-SNE) to increase the effectiveness of on-site inspection. Results: A number of simultaneous violation patterns was discovered by applying Association Rule Mining to FDA's inspection data collected from October 2008 to February 2022. Among them there were 'concurrent violation patterns' derived from similar regulatory ranges of two or more regulations. These patterns do not help to predict violations that simultaneously appear but belong to different regulations. Those unnecessary patterns were excluded by applying t-SNE based on text-mining. Conclusion: Our proposed approach enables the recognition of simultaneous violation patterns during the on-site inspection. It is expected to decrease the detection time by increasing the likelihood of finding intentionally concealed violations.

• KIISE Transactions on Computing Practices
• /
• v.23 no.4
• /
• pp.226-231
• /
• 2017

Word representation, an important area in natural language processing(NLP) used machine learning, is a method that represents a word not by text but by distinguishable symbol. Existing word embedding employed a large number of corpora to ensure that words are positioned nearby within text. However corpus-based word embedding needs several corpora because of the frequency of word occurrence and increased number of words. In this paper word embedding is done using dictionary definitions and semantic relationship information(hypernyms and antonyms). Words are trained using the feature mirror model(FMM), a modified Skip-Gram(Word2Vec). Sense similar words have similar vector. Furthermore, it was possible to distinguish vectors of antonym words.

• Journal of Advanced Navigation Technology
• /
• v.20 no.6
• /
• pp.617-624
• /
• 2016

The hangul word embedding should be performed certainly process for noun extraction. Otherwise, it should be trained words that are not necessary, and it can not be derived efficient embedding results. In this paper, we propose model that can retrieve more efficiently by query language expansion using hangul word embedded, apriori, and text mining. The word embedding and apriori is a step expanding query language by extracting association words according to meaning and context for query language. The hangul text mining is a step of extracting similar answer and responding to the user using noun extraction, TF-IDF, and cosine similarity. The proposed model can improve accuracy of answer by learning the answer of specific domain and expanding high correlation query language. As future research, it needs to extract more correlation query language by analysis of user queries stored in database.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.2
• /
• pp.137-142
• /
• 2020

Many studies based on I-vector have been conducted in a variety of environments, from text-dependent short-utterance to text-independent long-utterance. In this paper, we propose a speaker verification system employing a combination of I-vector with Probabilistic Linear Discriminant Analysis (PLDA) and speaker embedding of Long Short Term Memory (LSTM) with attention mechanism in far-field and noisy environments. The LSTM model's Equal Error Rate (EER) is 15.52 % and the Attention-LSTM model is 8.46 %, improving by 7.06 %. We show that the proposed method solves the problem of the existing extraction process which defines embedding as a heuristic. The EER of the I-vector/PLDA without combining is 6.18 % that shows the best performance. And combined with attention-LSTM based embedding is 2.57 % that is 3.61 % less than the baseline system, and which improves performance by 58.41 %.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2000.11a
• /
• pp.267-275
• /
• 2000

최근 전자상거래가 활성화되면서, 거래 인증서와 같은 온라인 증명서가 광범위하게 사용되고 있다. 그리고 증명서의 위/변조기술이 발전함에 따라서 온라인 거래에 사용되는 증명서의 인증과 위조/변조를 방지하는 기술이 필요하게 되었다. 본 연구는 증명서의 인증에 필요한 기술로서, 메시지 인증함수가 가지는 성질을 포함하고 있다. 본 연구에서 개발한 알고리즘은 증명서에 포함된 텍스트문서가 위조/변조되었을 경우 그 변동 상황을 알아내며, 부정적으로 위조/변조된 부분을 검출하며, 변동상황 검출과 함께 원 증명서의 문서를 복원할 수 있는 기술이다. 만일 이 증명서에 대하여 변동이 진행된 흔적이 발견될 경우, 증명서를 인증하지 않으며, 삽입한 텍스트 데이터를 추출하고 변동을 확인하는 것과 함께 필요한 정보를 복원한다. 본 논문의 시험결과에 근거하면 256$\times$256BMP file Format 이미지에 3만2천자 정도의 텍스트문서를 삽입할 수 있었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.307-310
• /
• 2020

본 논문에서 이슈 키워드 추출을 위해 텍스트 마이닝(Text Mining) 기술을 요구한다. 사회적 이슈 키워드를 추출하기 위해 키워드 수집 모델이 되는 사이트에서 크롤링(crawling)을 수행한 뒤, 형태소 단위 의미있는 단어를 수집하기 위해 형태소 분석(morphological analysis)을 수행한다. 한국어 형태소 분석을 위해 파이썬의 코엔엘파이(KoNLPy) 패키지를 활용한다. 형태소 분석을 통해 나뉘어진 단어에서 통계를 내어 이슈 키워드 추출한다. 이슈 키워드를 뒷받침할 연관 단어를 분석하기 위해 단어 임베딩(Word Embedding)을 수행한다. 단어 임베딩 수행을 위해 Word2Vec 모델 중 Skip-Gram 방법론을 적용하여 연관 단어를 분석하도록 개발하였다. 웹 소켓(Web Socket) 통신을 통한 채팅 프로그램의 상단에 분석한 이슈 키워드와 연관 단어를 출력하도록 개발하였다.

• Journal of information and communication convergence engineering
• /
• v.21 no.1
• /
• pp.9-16
• /
• 2023

The natural language on social network platforms has a certain front-to-back dependency in structure, and the direct conversion of Chinese text into a vector makes the dimensionality very high, thereby resulting in the low accuracy of existing text classification methods. To this end, this study establishes a deep learning model that combines a big data ultra-deep convolutional neural network (UDCNN) and long short-term memory network (LSTM). The deep structure of UDCNN is used to extract the features of text vector classification. The LSTM stores historical information to extract the context dependency of long texts, and word embedding is introduced to convert the text into low-dimensional vectors. Experiments are conducted on the social network platforms Sogou corpus and the University HowNet Chinese corpus. The research results show that compared with CNN + rand, LSTM, and other models, the neural network deep learning hybrid model can effectively improve the accuracy of text classification.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.2
• /
• pp.148-154
• /
• 2021

Text-Independent speaker verification needs to extract text-independent speaker embedding to improve generalization performance. However, deep neural networks that depend on training data have the potential to overfit text information instead of learning the speaker information when repeatedly learning from the identical time series. In this paper, to prevent the overfitting, we propose a segment unit shuffling layer that divides and rearranges the input layer or a hidden layer along the time axis, thus mixes the time series information. Since the segment unit shuffling layer can be applied not only to the input layer but also to the hidden layers, it can be used as generalization technique in the hidden layer, which is known to be effective compared to the generalization technique in the input layer, and can be applied simultaneously with data augmentation. In addition, the degree of distortion can be adjusted by adjusting the unit size of the segment. We observe that the performance of text-independent speaker verification is improved compared to the baseline when the proposed segment unit shuffling layer is applied.

• ETRI Journal
• /
• v.44 no.5
• /
• pp.794-804
• /
• 2022

Transformer models have had a great impact on natural language processing (NLP) in recent years by realizing outstanding and efficient contextualized language models. Recent studies have used transformer-based language models for various NLP tasks, including Persian named entity recognition (NER). However, in complex tasks, for example, NER, it is difficult to determine which contextualized embedding will produce the best representation for the tasks. Considering the lack of comparative studies to investigate the use of different contextualized pretrained models with sequence modeling classifiers, we conducted a comparative study about using different classifiers and embedding models. In this paper, we use different transformer-based language models tuned with different classifiers, and we evaluate these models on the Persian NER task. We perform a comparative analysis to assess the impact of text representation and text classification methods on Persian NER performance. We train and evaluate the models on three different Persian NER datasets, that is, MoNa, Peyma, and Arman. Experimental results demonstrate that XLM-R with a linear layer and conditional random field (CRF) layer exhibited the best performance. This model achieved phrase-based F-measures of 70.04, 86.37, and 79.25 and word-based F scores of 78, 84.02, and 89.73 on the MoNa, Peyma, and Arman datasets, respectively. These results represent state-of-the-art performance on the Persian NER task.