텍스트마이닝 방법론을 활용한 기업 부도 예측 연구

최정원 ( Jung Won Choi ),한호선 ( Ho Sun Han ),이미영 ( Mi Young Lee ),안준모 ( Jun Mo Ahn ) 한국생산성학회 2015 生産性論集 Vol.29 No.1

Traditional corporate bankruptcy prediction methodology basically relies on financial accounting data to objectively reflect the status of companies. However, since financial accounting data is difficult to immediately reflect changes in the status of companies, real-time financial data such as stock and bond prices are also used in order to make up for the shortcomings. In this study, we use news text information which is a typical real-time information to study the corporate bankruptcy prediction models. In the past, news text information was difficult to use in quantitative analysis but not any more due to the recent advances of information processing technology and text-mining techniques. For bankruptcy prediction using news information, we collect news text for six months before the bankruptcy events of companies actually occur and study the possibility of bankruptcy prediction based on the data by utilizing text-mining techniques. Results indicate that we can not get such a high level of predictability as that of existing corporate bankruptcy prediction models, but that there exists a high potential of this approach enough to increase the predictability of bankruptcy models. Further research on bankruptcy prediction model using news text information will be promising.

Improving a model for the text input time prediction

Yong-Gi Bae,Seung-Kweon Hong 대한인간공학회 2014 대한인간공학회 학술대회논문집 Vol.2014 No.11

Objective: The aim of this study is to suggest a revised model for prediction of text entry time. Background: Several studies suggested mathematical models to predict text entry times. However, the prediction models did not consider the level of typing skill. Method: This study suggested a revised model to predict text input times. The model consists of 4 sub-models. In order to evaluate the model, input times by typists with several skill levels were measured and were compared with predicted input times. Results: Prediction models for the high level of skill typists provided the shorter input time than real input time, while prediction models for the low level of skill typists provided the longer input time than real input time. Conclusion: The suggested models could not exactly predict text input times. However, a pattern of changing text input times was predicted. Application: The results of this study might help to guide the future study on text input time prediction.

소형정보기기를 위한 지능형 가상 키패드 입력에 관한 연구

유승헌(Yoo, Seunghun) 인제대학교 디자인연구소 2012 Journal of Integrated Design Research (JIDR) Vol.11 No.2

모바일 터치스크린 단말 상에서 텍스트 입력의 컨텍스트가 다양화 되면서, 작은 크기의 물리적 공간 제약에 다른 타이핑 효율 문제 해결 또한 중요한 논제가 되고 있다. 모바일 상에서 이를 위한 여러 지원 기능들이 있는데, 영미권에서 가장 많이 사용되는 단어예측(Predictive Texting)기능은 인지적 레벨(Cognition Level)의 가장 대표적인 지원기능이며, 예측된 단어가 의도와 맞을 때에는 키 스트로크 효율을 높이지만, 사용자의 의도와 다른 단어가 제시되는 경우가 많아 화면 영역 과다 차지, 동음이의어 선택 태스크 추가, 단어삭제 추가 수행 에러 등 효율저하가 발생하고 있다. 본 연구에서는 이 문제를 해결하기 위해, 예측된 단어가 잘못 예측된 상황을 잠재적 Noise로 규정하고, 예측된 단어를 한번에 제시하는 인지적 레벨 지원 대신, 예측된 단어의 다음 철자를 단계별로 한 글자씩 키패드를 강조해주는 수행레벨(operation level support)을 함께 지원하는 PCNC방식을 제안하였다. 제안된 디자인 방식은 스마트폰 상의 프로토타입으로 제작되어, 일반적인 예측 모드와 사용자들의 입력효율을 비교, 분석하였다. 동일한 문장을 입력할 시 사용자 수행도를 비교한 결과 PCNC는 일반 입력체계 및 T9체계 대비 정확도, 오타율, 속도 비율에서 통계적으로 유의한 개선율을 나타내었다. Text typing on touch screen of smart mobile devices is becoming more high frequency task. The goal and context of text input are developing broader while the limitation of physical size of operational area of mobile is affecting onto low efficiency of task compared to full PC experiences One of the mobile text input aiding solutions, predictive texting function, is supporting users with suggestion of word on display which is close to users’ intention based on prediction from parsed characters. It is generating cognition level aid and contributes to reduce many key stokes for users in case the prediction is right. In the other hand, it causes cognitive and operational error for users to correct in addition ? taking large part of screen, additional selection task of similar spell, erasing miss typo- when the suggested word does not match with users’ intention. PCNC system, suggested in this research defines the wrong predicted word as noise, and supports users with highlighting the available one-keypad in advance with the order of spell instead of showing the full word in the popup on the screen. The comparative experiments was conducted with the mobile based prototype of PCNC and current predictive functions. The measure the efficiency of typing were elapsed time and error occurrence for task performance. The result showed that PCNC supports better performance than predictive text, especially when PCNC is combined with predictive text.

한국어 영화평 댓글 자료 및 다항 역회귀 기법을 활용한 영화 평점 예측 분석

김현진,김지수,이은령 한국통계학회 2023 응용통계연구 Vol.36 No.3

With the development of IT technology, a vast amount of information, including articles, blogs and online comments, is being circulated in modern society. Much of this data consists of text, which is unstructured data and is constantly being produced. As a result, the demand for analysis using text mining continues to increase by converting unstructured data into new digitized data. Unlike data mining, which extracts patterns from the structured data, the biggest difference is that text mining looks for patterns from the unstructured text. A main purpose of this study is to examine how high-dimensional Korean text variables, e.g., Korean movie reviews, can be summarized to provide sufficient information and to reduce dimension. This helps a further analysis, e.g., prediction of movie ratings. We use a state of the art method called multinomial inverse regression (MNIR), which was proposed to obtain a low-dimensional summary of texts that contains sentiment information. This requires the review comments to convert into an appropriate form for the model, so it was preprocessed and quantified in various ways. Besides prediction evaluation resulting from the summarized information, we propose the multiple MNIR models that consider interaction terms with movie genre variables. The proposed methods reveal an interesting feature of text data and it show a better prediction performance in our empirical study. IT 기술의 발달으로 인해 현대 사회에는 온라인 상의 기사, 블로그, 댓글 등 방대한 정보가 유통되고 있다. 이러한 데이터의 대부분은 비정형적인 자료(unstructured data)인 텍스트로 구성되어 있으며 끊임없이 생산되고 있다. 이에 따라 비정형적인 자료를 새롭게 수치화된 자료로 변환하여 텍스트 마이닝을 활용한 분석의 수요가 지속적으로 증가하고 있다. 정형화된 자료로부터 패턴을 찾는 데이터 마이닝(data mining)과 달리 텍스트 마이닝은 이렇게 비정형화된 텍스트로부터 패턴을 찾는다는 점이 가장 큰 차이점이다. 본 연구에서는 고차원의 한국어 텍스트를 최소한의 저차원으로 충분한 정보를 담고 있는지 살펴보는 것을 목적으로 한다. 여러 선행 연구로부터 이미 차원을 축소하는 방법이 소개되었으며, 그 중 다항 역회귀 기법과 충분차원축소 개념을 활용한다. 다항 역회귀 모형 기법을 활용하기 위해서는 리뷰 댓글을 모형에 알맞은 형태로 변환해야 하기 때문에 다양한 방식으로 전처리하고 수치화하였다. 분석 과정에서 요약된 저차원의 정보로 생기는 예측력 등과 같은 문제점을 파악하며 이를 해결하기 위한 방법으로 교차항이 포함된 모형을 제안한다. 또한 차원 축소 방법을 이용하여 기존 모형과 비교 분석하고 최종으로 최소한의 차원으로 예측력과 설명력을 기술한다.

순행적 추론의 활성화 연구

유덕근 ( Duck Geun Yoo ) 한국외국어대학교 언어연구소 2010 언어와 언어학 Vol.0 No.49

This paper examines inference, i.e., the act of pursuing cohesion and coherence in text by activating implicit information based on knowledge stored in the memories. In particular, this paper aims to discuss the dichotomy between “bridging inferences” and “predictive inferences” based on theoretical and empirical approaches. In the empirical level, this paper introduces the results of experiment that the subject``s predictive inference process was investigated in real time by parameterizing logical links in Korean. The experiment results show that as SMF-Model assumes, when enough context and situation information is provided, predictive inferences can occur prior to propositional level and be included at the final representation of text.

부도예측 모형에서 뉴스 분류를 통한 효과적인 감성분석에 관한 연구

김찬송,신민수 한국IT서비스학회 2019 한국IT서비스학회지 Vol.18 No.1

Bankruptcy prediction model is an issue that has consistently interested in various fields. Recently, as technology for dealing with unstructured data has been developed, researches applied to business model prediction through text mining have been activated, and studies using this method are also increasing in bankruptcy prediction. Especially, it is actively trying to improve bankruptcy prediction by analyzing news data dealing with the external environment of the corporation. However, there has been a lack of study on which news is effective in bankruptcy prediction in real-time mass-produced news. The purpose of this study was to evaluate the high impact news on bankruptcy prediction. Therefore, we classify news according to type, collection period, and analyzed the impact on bankruptcy prediction based on sentiment analysis. As a result, artificial neural network was most effective among the algorithms used, and commentary news type was most effective in bankruptcy prediction. Column and straight type news were also significant, but photo type news was not significant. In the news by collection period, news for 4 months before the bankruptcy was most effective in bankruptcy prediction. In this study, we propose a news classification methods for sentiment analysis that is effective for bankruptcy prediction model.

헌법문언에 반하는 헌법해석

이황희 법조협회 2019 法曹 Vol.68 No.5

Since the constitutional interpretation is based on the constitutional text, the constitutional interpretation against the constitutional text may sound like a contradiction. However, this interpretation, though rarely, is applied to the constitutional cases and claimed as a doctrine. In this article, I have examined why this phenomenon occurs, what cases and arguments can be found in terms of this interpretation, and under which conditions this interpretation can be justified. Among cases of domestic and foreign Constitutional Courts and arguments in our academic society, there are those that can be regarded as the constitutional interpretation against the constitutional text. The interpretation of the Constitutional Court of Korea on whether foreigners can be the subject of basic rights, some scholars’ claims related with it, the interpretation of many foreign constitutional courts on the jurisdiction to review constitutional amendments, and some of the interpretations on the Article 8 of the current Constitution, and on the Article 53 paragraph 4 of the 1972 Constitution can be explained from this point of view. The problem of the legal norm which is legitimate formally, but illegitimate materially is generally said to be resolved with the notion of the hierachy of norms with the limited government, and the constitutional review that fulfills the supremacy of the Constitution. However, since the Constitution doesn’t have norms superior to it, this problem is still latent at the level of the Constitution. The constitutional interpretation against the constitutional text is considered as a solution to this exceptional problem. The specific provisions of the Constitution are not always right and just. However, such provisions don’t always justify the constitutional interpretation against it. This kind of constitutional interpretation can be taken into account in a situation where the basic principle and philosophy of the Constitution are seriously undermined when the constitutional text is applied as written. When the problem caused by it may be resolved by a rapid constitutional amendment or by harmonious interpretation, namely practical concordance, it is desirable that this interpretation be restrained. 헌법해석은 법문언을 기초로 하므로, 문언에 반하는 헌법해석은 모순적으로 들린다. 그러나 드물기는 하지만 이 해석기법은 재판에 활용되고 학설로도 주장된다. 본고에서는 왜 이러한 현상이 생겨나는지, 그러한 해석사례들로는 어떤 것들이 있고 또 이것이 실제로 존재하는 해석이라면 과연 어떤 조건하에서 정당화될 수 있는지를 살펴보았다. 국내외의 사례들과 우리 학계에서의 주장들 중에는 문언에 반하는 헌법해석으로 볼 수 있는 것들이 있다. 그 중에서 외국인의 기본권주체성에 관한 우리 헌법재판소의 해석, 그와 관련된 일부 학설들, 헌법개정의 위헌여부 심사권에 관한 여러 해외 헌법재판소들의 해석, 그리고 현행헌법 제8조 제4항과 유신헌법 제53조 제4항에 관한 일부 학설들의 해석을 짚어 보았다. 형식적으로는 정당하지만 내용적으로는 부당한 법의 문제는 헌법을 최고규범으로 하는 규범서열과 헌법에 의한 국가권력의 구속과 제한을 추구하는 제한정부 관념, 그리고 이를 관철할 수 있는 위헌법률심사 제도를 통해 성공적으로 해결되었다고 평가된다. 그러나 헌법은 그보다 더 우월한 규범이 없기 때문에, 헌법의 차원에서는 이 문제가 여전히 잠재되어 있다. 문언에 반하는 헌법해석은 이 예외적인 상황에 대한 해법으로 고려될 수 있다. 헌법의 개별 규정이 늘 옳고 정의로운 것은 아니지만, 헌법에 부당한 내용이 포함되어 있다 해서 문언에 반하는 헌법해석이 언제나 정당화되는 것도 아니다. 그 조항이 헌법의 기본원칙과 이념을 심각하게 훼손해 헌법의 문언에서 벗어나는 해석만이 헌법의 기본원칙과 이념을 보호할 수 있는 경우에 한해 그러한 해석이 예외적으로 정당화될 수 있다. 그 조항으로 인한 문제가 조속한 개헌을 통해 혹은 조화적 해석을 통해 해결될 가능성이 있을 때 이러한 해석은 자제되어야 한다.

온라인 리뷰의 텍스트 분석을 통한 고객 평점 예측에 관한 연구

윤민지(Minji Yoon),조창규(Chang Kyoo Cho),서용원(Yong Won Seo) 한국생산관리학회 2021 한국생산관리학회지 Vol.32 No.1

최근 온라인 리뷰의 중요성이 강조됨에 따라, 온라인 리뷰와 고객 평점의 관계로부터 제품 및 서비스의 품질속성을 도출하는 방안에 대한 연구가 이루어지고 있다. 그러나, 많은 온라인 리뷰에서는 일관된 기준의 고객 평점이 존재하지 않아 평점 기반의 방법이 적용되기 어렵다. 이에 따라 본 연구에서는 텍스트 분석을 활용하여 고객 평점에 의존하지 않고 Kano품질속성과 극성 점수에 기반한 고객 평점 예측 모형을 제시하였다. 이를 통해 얻어진 평점의 실제 온라인 리뷰에 대한 결정계수가 유의미하게 나타남을 확인하여, 본 연구에서 제시한 Kano 품질속성을 기반으로 한 고객 평점 예측 모형의 유효성을 입증하였다. 나아가 동일한 Kano 품질속성을 가지는 호텔 성급을 그룹화 하여 각 그룹별로 Kano 품질속성의 특성을 고려하는 경우에 예측력이 더욱 높아짐을 확인하여, 차별화된 Kano 품질속성을 적극적으로 평점 예측에 고려하는 것이 예측력을 높이는 데에 기여할 수 있음을 확인하였다. 이러한 결과는 평점이 명시적으로 나타나 있지 않은 고객 리뷰에서 고객들이 잠재적으로 평가하는 평점을 정량화할 수 있음을 입증하여 실용적 의미가 크다. The usefulness of online reviews in understanding customers’ perceptions of quality satisfaction attributes has prompted a growing number of studies to present ways to predict customer ratings based on online review texts. However, these studies rely on the relationship between review texts and customer ratings, which cannot be applied to reviews without explicitly stated rating values. In this study, we propose a model to predict customer ratings from online review texts without explicit customer ratings. We applied text analysis to hotel review texts collected on the internet and extracted the Kano quality attributes of hotel services, based on which we established a model to predict potential customer ratings. Confirming that the coefficient of determination for the actual online reviews of the ratings obtained with this model is significant, we demonstrated the validity of the prediction model for customer ratings based on the Kano quality attributes. Furthermore, after grouping hotel star grades having identical Kano quality attributes, we confirmed that predictive power is further enhanced when the Kano quality attributes differentiated by group are considered. Thus, we confirmed that actively considering the specialized Kano quality attributes in the prediction of ratings contributes to enhancing the predictive power. These findings suggest that it is possible to quantify the ratings customers award in their reviews where rating scores are not explicitly shown, which indicates substantial significance for practical use.

헌법문언에 반하는 헌법해석

李煌熙 ( Hwanghee Lee ) 법조협회 2019 法曹 Vol.68 No.5

헌법해석은 법문언을 기초로 하므로, 문언에 반하는 헌법해석은 모순적으로 들린다. 그러나 드물기는 하지만 이 해석기법은 재판에 활용되고 학설로도 주장된다. 본고에서는 왜 이러한 현상이 생겨나는지, 그러한 해석사례들로는 어떤 것들이 있고 또 이것이 실제로 존재하는 해석이라면 과연 어떤 조건하에서 정당화될 수 있는지를 살펴보았다. 국내외의 사례들과 우리 학계에서의 주장들 중에는 문언에 반하는 헌법해석으로 볼 수 있는 것들이 있다. 그 중에서 외국인의 기본권주체성에 관한 우리 헌법재판소의 해석, 그와 관련된 일부 학설들, 헌법개정의 위헌여부 심사권에 관한 여러 해외 헌법재판소들의 해석, 그리고 현행헌법 제8조 제4항과 유신헌법 제53조 제4항에 관한 일부 학설들의 해석을 짚어 보았다. 형식적으로는 정당하지만 내용적으로는 부당한 법의 문제는 헌법을 최고규범으로 하는 규범서열과 헌법에 의한 국가권력의 구속과 제한을 추구하는 제한정부 관념, 그리고 이를 관철할 수 있는 위헌법률심사 제도를 통해 성공적으로 해결되었다고 평가된다. 그러나 헌법은 그보다 더 우월한 규범이 없기 때문에, 헌법의 차원에서는 이 문제가 여전히 잠재되어 있다. 문언에 반하는 헌법해석은 이 예외적인 상황에 대한 해법으로 고려될 수 있다. 헌법의 개별 규정이 늘 옳고 정의로운 것은 아니지만, 헌법에 부당한 내용이 포함되어 있다 해서 문언에 반하는 헌법해석이 언제나 정당화되는 것도 아니다. 그 조항이 헌법의 기본원칙과 이념을 심각하게 훼손해 헌법의 문언에서 벗어나는 해석만이 헌법의 기본원칙과 이념을 보호할 수 있는 경우에 한해 그러한 해석이 예외적으로 정당화될 수 있다. 그 조항으로 인한 문제가 조속한 개헌을 통해 혹은 조화적 해석을 통해 해결될 가능성이 있을 때 이러한 해석은 자제되어야 한다. Since the constitutional interpretation is based on the constitutional text, the constitutional interpretation against the constitutional text may sound like a contradiction. However, this interpretation, though rarely, is applied to the constitutional cases and claimed as a doctrine. In this article, I have examined why this phenomenon occurs, what cases and arguments can be found in terms of this interpretation, and under which conditions this interpretation can be justified. Among cases of domestic and foreign Constitutional Courts and arguments in our academic society, there are those that can be regarded as the constitutional interpretation against the constitutional text. The interpretation of the Constitutional Court of Korea on whether foreigners can be the subject of basic rights, some scholars’ claims related with it, the interpretation of many foreign constitutional courts on the jurisdiction to review constitutional amendments, and some of the interpretations on the Article 8 of the current Constitution, and on the Article 53 paragraph 4 of the 1972 Constitution can be explained from this point of view. The problem of the legal norm which is legitimate formally, but illegitimate materially is generally said to be resolved with the notion of the hierachy of norms with the limited government, and the constitutional review that fulfills the supremacy of the Constitution. However, since the Constitution doesn’t have norms superior to it, this problem is still latent at the level of the Constitution. The constitutional interpretation against the constitutional text is considered as a solution to this exceptional problem. The specific provisions of the Constitution are not always right and just. However, such provisions don’t always justify the constitutional interpretation against it. This kind of constitutional interpretation can be taken into account in a situation where the basic principle and philosophy of the Constitution are seriously undermined when the constitutional text is applied as written. When the problem caused by it may be resolved by a rapid constitutional amendment or by harmonious interpretation, namely practical concordance, it is desirable that this interpretation be restrained.

온라인 상품평의 내용적 특성이 소비자의 인지된 유용성에 미치는 영향

박윤주(Yoon-Joo Park),김경재(Kyoung-jae Kim) 한국지능정보시스템학회 2017 지능정보연구 Vol.23 No.3

In Internet commerce, consumers are heavily influenced by product reviews written by other users who have already purchased the product. However, as the product reviews accumulate, it takes a lot of time and effort for consumers to individually check the massive number of product reviews. Moreover, product reviews that are written carelessly actually inconvenience consumers. Thus many online vendors provide mechanisms to identify reviews that customers perceive as most helpful (Cao et al. 2011; Mudambi and Schuff 2010). For example, some online retailers, such as Amazon.com and TripAdvisor, allow users to rate the helpfulness of each review, and use this feedback information to rank and re-order them. However, many reviews have only a few feedbacks or no feedback at all, thus making it hard to identify their helpfulness. Also, it takes time to accumulate feedbacks, thus the newly authored reviews do not have enough ones. For example, only 20% of the reviews in Amazon Review Dataset (Mcauley and Leskovec, 2013) have more than 5 reviews (Yan et al, 2014). The purpose of this study is to analyze the factors affecting the usefulness of online product reviews and to derive a forecasting model that selectively provides product reviews that can be helpful to consumers. In order to do this, we extracted the various linguistic, psychological, and perceptual elements included in product reviews by using text-mining techniques and identifying the determinants among these elements that affect the usability of product reviews. In particular, considering that the characteristics of the product reviews and determinants of usability for apparel products (which are experiential products) and electronic products (which are search goods) can differ, the characteristics of the product reviews were compared within each product group and the determinants were established for each. This study used 7,498 apparel product reviews and 106,962 electronic product reviews from Amazon.com. In order to understand a review text, we first extract linguistic and psychological characteristics from review texts such as a word count, the level of emotional tone and analytical thinking embedded in review text using widely adopted text analysis software LIWC (Linguistic Inquiry and Word Count). After then, we explore the descriptive statistics of review text for each category and statistically compare their differences using t-test. Lastly, we regression analysis using the data mining software RapidMiner to find out determinant factors. As a result of comparing and analyzing product review characteristics of electronic products and apparel products, it was found that reviewers used more words as well as longer sentences when writing product reviews for electronic products. As for the content characteristics of the product reviews, it was found that these reviews included many analytic words, carried more clout, and related to the cognitive processes (CogProc) more so than the apparel product reviews, in addition to including many words expressing negative emotions (NegEmo). On the other hand, the apparel product reviews included more personal, authentic, positive emotions (PosEmo) and perceptual processes (Percept) compared to the electronic product reviews. Next, we analyzed the determinants toward the usefulness of the product reviews between the two product groups. As a result, it was found that product reviews with high product ratings from reviewers in both product groups that were perceived as being useful contained a larger number of total words, many expressions involving perceptual processes, and fewer negative emotions. In addition, apparel product reviews with a large number of comparative expressions, a low expertise index, and concise content with fewer words in each sentence were perceived to be useful. In the case of electronic product reviews, those that were analytical with a high expertise index, along with containing many authe