• Annual Conference on Human and Language Technology
• /
• 2013.10a
• /
• pp.64-69
• /
• 2013

스마트폰의 대중화로 인하여 앱 마켓 시장이 급속도로 성장하였다. 이로 인하여 하루에도 수십개의 새로운 앱들이 출시되고 있다. 이러한 앱 마켓 시장의 급격한 성장으로 인해 사용자들은 자신이 흥미를 가질만한 앱들을 선택하는데 큰 어려움을 겪고 있어 앱 추천 방법에 대한 연구에 많은 관심이 집중되고 있다. 기존 연구에서 협력 필터링 기반의 추천 방법들을 제안하였으나 이는 콜드 스타트 문제를 지니고 있다. 이와는 달리 컨텐츠 기반 필터링 방식은 콜드 스타트 문제를 효율적으로 해소할 수 있는 방법이지만 앱설명에는 광고, 공지사항등 실질적으로 앱의 특징과는 무관한 노이즈들이 다수 존재하고 이들은 앱 사이의 유사관계를 파악하는데 방해가 된다. 본 논문에서는 이런 문제를 해결하기 위하여 앱 설명에서 노이즈에 해당하는 설명들을 자동으로 제거할 수 있는 모델을 제안한다. 제안하는 모델은 모바일 앱 설명을 구성하고 있는 각 문단을 LDA로 학습된 토픽들의 비율로 나타내고 이들을 분류문제에서 우수한 성능을 보이는 SVM을 이용하여 분류한다. 실험 결과에 따르면 본 논문에서 제안한 방법은 기존에 문서 분류에 많이 사용되는 Bag-of-Word 표현법에 기반한 문서 표현 방식보다 더 나은 분류 성능을 보였다.

• Journal of Intelligence and Information Systems
• /
• v.19 no.3
• /
• pp.1-23
• /
• 2013

To discover significant social issues such as unemployment, economy crisis, social welfare etc. that are urgent issues to be solved in a modern society, in the existing approach, researchers usually collect opinions from professional experts and scholars through either online or offline surveys. However, such a method does not seem to be effective from time to time. As usual, due to the problem of expense, a large number of survey replies are seldom gathered. In some cases, it is also hard to find out professional persons dealing with specific social issues. Thus, the sample set is often small and may have some bias. Furthermore, regarding a social issue, several experts may make totally different conclusions because each expert has his subjective point of view and different background. In this case, it is considerably hard to figure out what current social issues are and which social issues are really important. To surmount the shortcomings of the current approach, in this paper, we develop a prototype system that semi-automatically detects social issue keywords representing social issues and problems from about 1.3 million news articles issued by about 10 major domestic presses in Korea from June 2009 until July 2012. Our proposed system consists of (1) collecting and extracting texts from the collected news articles, (2) identifying only news articles related to social issues, (3) analyzing the lexical items of Korean sentences, (4) finding a set of topics regarding social keywords over time based on probabilistic topic modeling, (5) matching relevant paragraphs to a given topic, and (6) visualizing social keywords for easy understanding. In particular, we propose a novel matching algorithm relying on generative models. The goal of our proposed matching algorithm is to best match paragraphs to each topic. Technically, using a topic model such as Latent Dirichlet Allocation (LDA), we can obtain a set of topics, each of which has relevant terms and their probability values. In our problem, given a set of text documents (e.g., news articles), LDA shows a set of topic clusters, and then each topic cluster is labeled by human annotators, where each topic label stands for a social keyword. For example, suppose there is a topic (e.g., Topic1 = {(unemployment, 0.4), (layoff, 0.3), (business, 0.3)}) and then a human annotator labels "Unemployment Problem" on Topic1. In this example, it is non-trivial to understand what happened to the unemployment problem in our society. In other words, taking a look at only social keywords, we have no idea of the detailed events occurring in our society. To tackle this matter, we develop the matching algorithm that computes the probability value of a paragraph given a topic, relying on (i) topic terms and (ii) their probability values. For instance, given a set of text documents, we segment each text document to paragraphs. In the meantime, using LDA, we can extract a set of topics from the text documents. Based on our matching process, each paragraph is assigned to a topic, indicating that the paragraph best matches the topic. Finally, each topic has several best matched paragraphs. Furthermore, assuming there are a topic (e.g., Unemployment Problem) and the best matched paragraph (e.g., Up to 300 workers lost their jobs in XXX company at Seoul). In this case, we can grasp the detailed information of the social keyword such as "300 workers", "unemployment", "XXX company", and "Seoul". In addition, our system visualizes social keywords over time. Therefore, through our matching process and keyword visualization, most researchers will be able to detect social issues easily and quickly. Through this prototype system, we have detected various social issues appearing in our society and also showed effectiveness of our proposed methods according to our experimental results. Note that you can also use our proof-of-concept system in http://dslab.snu.ac.kr/demo.html.

• Journal of Korea Multimedia Society
• /
• v.23 no.7
• /
• pp.883-890
• /
• 2020

In this paper, we propose a variational expectation-maximization algorithm that computes posterior probabilities from Latent Dirichlet Allocation (LDA) model. The algorithm approximates the intractable posterior distribution of a document term matrix generated from a corpus made up by 50 papers. It approximates the posterior by searching the local optima using lower bound of the true posterior distribution. Moreover, it maximizes the lower bound of the log-likelihood of the true posterior by minimizing the relative entropy of the prior and the posterior distribution known as KL-Divergence. The experimental results indicate that documents clustered to image classification and segmentation are correlated at 0.79 while those clustered to object detection and image segmentation are highly correlated at 0.96. The proposed variational inference algorithm performs efficiently and faster than Gibbs sampling at a computational time of 0.029s.

• 한국체육학회지인문사회과학편
• /
• v.57 no.2
• /
• pp.279-289
• /
• 2018

The purpose of this study is to investigate features of leisure and recreation scholarship study in The Korean Journal of physical education based on co-authors network and topic modeling through using Word Cloud and LDA Topic Modeling(Latent Dirichlet Allocation). The data collected for this study are 2,697 papers published online from January 2008 to March 2017 on the Korean journal of physical education. Respectively ordered analysis targets are the major author, author of correspondence, co-author 1, co-author 2, co-author n in related document to explore studies' trends using the 369 documents. As a result, the co-author network analysis result found that 451 were linked to the research network, on average researchers had 1.52 relationships and the average distance between researchers was 2.33. The Representative author's concentration of connection was ranked high in the order of the following, Lee. K. M., Hwang. S. H., H., Lee. C. S., and proximity centers were shown in Seo K. B., Han. J. H., Kim. K. J. Finally, parameter-centric features appeared in order of Lee. C. W. and Seo. K. B. was most actively connected between the researchers of the leisure-related academic papers. Future research needs discussions among scholars regarding the trend and direction of future leisure research.

• Economic and Environmental Geology
• /
• v.54 no.3
• /
• pp.353-364
• /
• 2021

Since the mid-twentieth century, geology has gradually evolved as an interdisciplinary context in South Korea. The journal of Economic and Environmental Geology (EEG) has a long history of over 52 years and published interdisciplinary articles based on geology. In this study, we performed a literature review using topic modeling based on Latent Dirichlet Allocation (LDA), an unsupervised machine learning model, to identify geological topics, historical trends (classic topics and emerging topics), and association by analyzing titles, keywords, and abstracts of 2,571 publications in EEG during 1968-2020. The results showed that 8 topics ('petrology and geochemistry', 'hydrology and hydrogeology', 'economic geology', 'volcanology', 'soil contaminant and remediation', 'general and structural geology', 'geophysics and geophysical exploration', and 'clay mineral') were identified in the EEG. Before 1994, classic topics ('economic geology', 'volcanology', and 'general and structure geology') were dominant research trends. After 1994, emerging topics ('hydrology and hydrogeology', 'soil contaminant and remediation', 'clay mineral') have arisen, and its portion has gradually increased. The result of association analysis showed that EEG tends to be more comprehensive based on 'economic geology'. Our results provide understanding of how geological research topics branch out and merge with other fields using a useful literature review tool for geological research in South Korea.

• Journal of Korean Library and Information Science Society
• /
• v.53 no.3
• /
• pp.455-477
• /
• 2022

This study aimed to compare and analyze librarian teacher evaluation index from evaluation of teachers' competency development with the the topics of information utilization education. To this end, LDA topic modeling was conducted by collecting papers related to information utilization education published in four major journals in the field of literature and information from 1995 to May 2022. As a result of topic modeling, it can be seen that information utilization education (T10) was the most actively discussed at 12.0% of the 20 topics, followed by library utilization classes (T2) 10.4% and user service (T3) 8.8%.On the other hand, 3.3% of reading discussion (T7), 2.9% of reading education (T19), 2.1% of manpower management (T13), and 2.1% of librarian teacher job satisfaction (T17) showed the lowest distributions 3.3%, 2.9%, 2.1%, and 2.1%, respectively. In addition, although librarian teacher's class model development (T1) and curriculum development (T20) are essential processes for collaborative classes and information utilization education, they were not reflected in the current teacher competency development evaluation index. Therefore, this study proposed that 'instructional model and curriculum development' indicator should be added on 'training and support classes' factors in the Librarian Teacher Evaluation Index in Teachers' Competency Development Evaluation for further evaluation.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.4
• /
• pp.57-70
• /
• 2021

This study aims to investigate the security demand about the traffic policing by analyzing civil complaints. Latent Dirichlet Allocation(LDA) was applied to extract key topics for 2,062 civil complaints data related to traffic policing from e-People. And additional analysis was made of reports of violations, which accounted for a high proportion. In this process, the consistency and convergence of keywords and representative documents were considered together. As a result of the analysis, complaints related to traffic police could be classified into 41 topics, including traffic safety facilities, passing through intersections(signals), provisional impoundment of vehicle plate, and personal mobility. It is necessary to strengthen crackdowns on violations at intersections and violations of motorcycles and take preemptive measures for the installation and operation of unmanned traffic control equipments, crosswalks, and traffic lights. In addition, it is necessary to publicize the recently amended laws a implemented policies, e-fine, procedure after crackdown.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.12
• /
• pp.261-270
• /
• 2020

In this study, we identify promising technologies for Piggyback transportation system by analyzing the relevant patent information. In order for this, we first develop the patent database by extracting relevant technology keywords from the pioneering research papers for the Piggyback flactcar system. We then employed textmining to identify the frequently referred words from the patent database, and using these words, we applied the LDA (Latent Dirichlet Allocation) algorithm in order to identify "topics" that are corresponding to "key" technologies for the Piggyback system. Finally, we employ the ARIMA model to forecast the trends of these "key" technologies for technology forecasting, and identify the promising technologies for the Piggyback system. with keyword search method the patent analysis. The results show that data-driven integrated management system, operation planning system and special cargo (especially fluid and gas) handling/storage technologies are identified to be the "key" promising technolgies for the future of the Piggyback system, and data reception/analysis techniques must be developed in order to improve the system performance. The proposed procedure and analysis method provides useful insights to develop the R&D strategy and the technology roadmap for the Piggyback system.

• The Journal of Information Systems
• /
• v.30 no.4
• /
• pp.277-301
• /
• 2021

Purpose The purpose of this study is to analyze the sentiment responses of the general public to non-face-to-face work using text mining methodology. As the number of non-face-to-face complaints is increasing over time, it is difficult to review and analyze in traditional methods such as surveys, and there is a limit to reflect real-time issues. Approach This study has proposed a method of the research model, first by collecting and cleansing the data related to non-face-to-face work among tweets posted on Twitter. Second, topics and keywords are extracted from tweets using LDA(Latent Dirichlet Allocation), a topic modeling technique, and changes for each section are analyzed through DTM(Dynamic Topic Modeling). Third, the complaints of non-face-to-face work are analyzed through the classification of positive and negative polarity in the COVID-19 section. Findings As a result of analyzing 1.54 million tweets related to non-face-to-face work, the number of IDs using non-face-to-face work-related words increased 7.2 times and the number of tweets increased 4.8 times after COVID-19. The top frequently used words related to non-face-to-face work appeared in the order of remote jobs, cybersecurity, technical jobs, productivity, and software. The words that have increased after the COVID-19 were concerned about lockdown and dismissal, and business transformation and also mentioned as to secure business continuity and virtual workplace. New Normal was newly mentioned as a new standard. Negative opinions found to be increased in the early stages of COVID-19 from 34% to 43%, and then stabilized again to 36% through non-face-to-face work sentiment analysis. The complaints were, policies such as strengthening cybersecurity, activating communication to improve work productivity, and diversifying work spaces.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.1
• /
• pp.29-42
• /
• 2019

In order to expand the qualitative aspects of public facility, this study used SNS data to derive user-oriented preference factors for public facilities and then were quantified in terms of supply side and demand side. To derive preference factor, LDA, one of topic modeling, was used and attraction index was calculated for each facility. In addition we analyzed spatial accessibility to measure the degree of service experience of users by using 2SFCA model. The study area covered public libraries of Seoul, Korea. As a result of study, five topics were extracted as preference factors for the public library: Circumstance, Scale of facility, Cultural program, Parenting, Books and materials. In particular topic of circumstance and parenting were newly derived preference factors unknown in previous studies. As a result of calculating attraction index for each library, the index of Songpa Library, Jungdok Library, and Namsan Library was high. Songpa library has received good evaluation in parenting factor, and Jungdok & Namsan library in circumstance factor. The accessibility of each region seems to better in center of Seoul where public libraries are crowded, but shrinking toward the outskirts. We expect that the proposed method will contribute to user-oriented public facility evaluation and policy decision making.