• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.7
• /
• pp.259-266
• /
• 2018

In this paper, we propose a method to classify a document using a Recurrent Neural Network by extracting features considering word sense and contexts. Word2vec method is adopted to include the order and meaning of the words expressing the word in the document as a vector. Doc2vec is applied for considering the context to extract the feature of the document. RNN classifier, which includes the output of the previous node as the input of the next node, is used as the document classification method. RNN classifier presents good performance for document classification because it is suitable for sequence data among neural network classifiers. We applied GRU (Gated Recurrent Unit) model which solves the vanishing gradient problem of RNN. It also reduces computation speed. We used one Hangul document set and two English document sets for the experiments and GRU based document classifier improves performance by about 3.5% compared to CNN based document classifier.

• Journal of Digital Convergence
• /
• v.18 no.1
• /
• pp.301-307
• /
• 2020

In recent years, a recommendation system is introduced in many different fields with the beginning of the 5G era and making a considerably prominent appearance mainly in books, movies, and music. In such a recommendation system, however, the preference degrees of users are subjective and uncertain, which means that it is difficult to provide accurate recommendation service. There should be huge amounts of learning data and more accurate estimation technologies in order to improve the performance of a recommendation system. Trying to solve this problem, this study proposed a movie recommendation system based on Fuzzy-AHP and Word2vec. The proposed system used Fuzzy-AHP to make objective predictions about user preference and Word2vec to classify scraped data. The performance of the system was assessed by measuring the accuracy of Word2vec outcomes based on grid search and comparing movie ratings predicted by the system with those by the audience. The results show that the optimal accuracy of cross validation was 91.4%, which means excellent performance. The differences in move ratings between the system and the audience were compared with the Fuzzy-AHP system, and it was superior at approximately 10%.

• Journal of the Korean Society for Library and Information Science
• /
• v.53 no.3
• /
• pp.247-271
• /
• 2019

Uncertainty in scientific knowledge means an uncertain state where propositions are neither true or false at present. The existing studies have analyzed the propositions written in the academic literature, and have conducted the performance evaluation based on the rule based and machine learning based approaches by using the corpus. Although they recognized that the importance of word construction, there are insufficient attempts to expand the word by analyzing the meaning of uncertainty words. On the other hand, studies for analyzing the structure of networks by using bibliometrics and text mining techniques are widely used as methods for understanding intellectual structure and relationship in various disciplines. Therefore, in this study, semantic relations were analyzed by applying Word2Vec to existing uncertainty words. In addition, WordNet, which is an English vocabulary database and thesaurus, was applied to perform a network analysis based on hypernyms, hyponyms, and synonyms relations linked to uncertainty words. The semantic and lexical relationships of uncertainty words were structurally identified. As a result, we identified the possibility of automatically expanding uncertainty words.

• International Journal of Computer Science & Network Security
• /
• v.22 no.8
• /
• pp.399-403
• /
• 2022

Natural texts are analyzed to obtain their intended meaning to be classified depending on the problem under study. One way to represent words is by generating vectors of real values to encode the meaning; this is called word embedding. Similarities between word representations are measured to identify text class. Word embeddings can be created using word2vec technique. However, recently fastText was implemented to provide better results when it is used with classifiers. In this paper, we will study the performance of well-known classifiers when using both techniques for word embedding with Arabic dataset. We applied them to real data collected from Wikipedia, and we found that both word2vec and fastText had similar accuracy with all used classifiers.

• The Journal of the Korea Contents Association
• /
• v.16 no.10
• /
• pp.687-692
• /
• 2016

Conventional way to search documents is keyword-based queries using vector space model, like tf-idf. Searching process of documents which is based on keywords can make some problems. it cannot recogize the difference of lexically different but semantically same words. This paper studies a scheme of document search based on document queries. In particular, it uses centrality vectors, instead of tf-idf vectors, to represent query documents, combined with the Word2vec method to capture the semantic similarity in contained words. This scheme improves the performance of document search and provides a way to find documents not only lexically, but semantically close to a query document.

• The Journal of Bigdata
• /
• v.3 no.1
• /
• pp.13-20
• /
• 2018

Stock market prediction has been long dream for researchers as well as the public. Forecasting ever-changing stock market, though, proved a Herculean task. This study proposes a novel stock market sentiment lexicon acquisition system that can predict the growth (or decline) of stock market index, based on economic news. For this purpose, we have collected 3-year's economic news from January 2015 to December 2017 and adopted Word2Vec model to consider the context of words. To evaluate the result, we performed sentiment analysis to collected news data with the automated constructed lexicon and compared with closings of the KOSPI (Korea Composite Stock Price Index), the South Korean stock market index based on economic news.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.795-796
• /
• 2021

본 논문은 국내 스타트업의 상품-서비스에 적합한 해외 바이어를 찾아 맟춤형으로 추천해주는 시스템을 설계하고자 한다. 추천 알고리즘은 CNN 기반의 Word2Vec과 Doc2Vec 알고리즘을 활용하며, 정확도를 높이기 위해 시각정보를 활용한다. 추천 시스템에 사용되는 데이터는 비정형 데이타인 회사 소개 및 상품/서비스 소개 문장 데이터이며, 제품 사진을 시각정보로 이용한다. 유사도가 높은 순으로 추천하기 위해 문장데이타를 키워드 리스트로 변환하고, Word2vec 모델에 이식시켜 키워드 좌표를 만들어 벡터화한다. 그리고, 문장의 중심점간 거리를 계산해 기업간 유사성 및 연관성을 도출한다. 이를 바탕으로 국내 스타트업의 문장데이타 및 시각정보와 유사도가 높은 순으로 해외바이어를 추천한다.

• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.307-309
• /
• 2016

이중언어 사전은 자연어처리 분야에서 매우 유용한 자원으로 사용되고 있다. 그러나 초기사전이나 병렬말뭉치 등 자원이 부족한 언어 쌍에 대해서 이중언어 사전을 추출하는 것은 쉽지 않다. 이러한 문제를 해결하기 위해 본 논문에서는 중간 언어 기반으로 Word2Vec와 CCA를 이용하여 이중언어 사전을 추출하는 방법을 제안한다. 본 논문에서 제안하는 방법의 성능을 평가하기 위해서 중간언어로 영어를 사용하여 스페인어-한국어에 대한 이중언어 사전을 추출하는 실험을 하였다. 무작위로 뽑은 200개의 단어에 대한 번역 정확도를 구하였다. 그 결과 최상위에서 37.5%, 상위 10위에서 63%, 그리고 상위 20위에서는 69.5%의 정확도를 얻을 수 있었다.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.307-309
• /
• 2016

이중언어 사전은 자연어처리 분야에서 매우 유용한 자원으로 사용되고 있다. 그러나 초기사전이나 병렬말뭉치 등 자원이 부족한 언어 쌍에 대해서 이중언어 사전을 추출하는 것은 쉽지 않다. 이러한 문제를 해결하기 위해 본 논문에서는 중간 언어 기반으로 Word2Vec와 CCA를 이용하여 이중언어 사전을 추출하는 방법을 제안한다. 본 논문에서 제안하는 방법의 성능을 평가하기 위해서 중간언어로 영어를 사용하여 스페인어-한국어에 대한 이중언어 사전을 추출하는 실험을 하였다. 무작위로 뽑은 200개의 단어에 대한 번역 정확도를 구하였다. 그 결과 최상위에서 37.5%, 상위 10위에서 63%, 그리고 상위 20위에서는 69.5%의 정확도를 얻을 수 있었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.11-12
• /
• 2021

소비자가 남긴 댓글이나 상품평은 상품기획의 주요 정보가 될 수 있다. 본 논문에서는 버티컬 무소음 마우스 7,300개에 대한 온라인 댓글을 딥러닝 기술인 Word2Vec을 이용하여 유사도 분석을 시행하였다. 유사도 분석결과 클릭 키워드에 대한 장점으로 소리(.975), 버튼(.972), 무게(.971)가 분석되었으며 단점은 가볍다(.959)이었다. 이는 구매 상품에 대한 소비자의 의견, 태도, 성향 및 서비스에 대한 포괄적인 의견들을 데이터화 하여 상품의 특징을 분석할 수 있는 의미있는 과정 이라고 볼 수 있다. 상품기획 프로세스에 딥러닝 기술을 통한 소비자의 감성분석자료 포함시키는 전략을 적용한다면 상품기획의 시간과 비용투자의 경제성을 높일 수 있고 나아가 빠르게 변화하는 소비자의 요구사항을 적기에 반영할 수 있을 것으로 생각된다.