삼각분할표 자료에서 베이지안 모형을 이용한 예측

이주미,임요한,한규섭,이경은,Lee, Ju-Mi,Lim, Jo-Han,Hahn, Kyu-S.,Lee, Kyeong-Eun 한국데이터정보과학회 2009 한국데이터정보과학회지 Vol.20 No.2

In the current paper, by extending Verall (1990)'s work, we propose a new Bayesian model for analyzing run-off triangle data. While Verall's (1990) work only account for the calendar year and evolvement time effects, our model further accounts for the "absolute time" effects. We also suggest a Markov Chain Monte Carlo method that can be used for estimating the proposed model. We apply our proposed method to analyzing three empirical examples. The results demonstrate that our method significantly reduces prediction error when compared with the existing methods. 본 논문은 삼각 분할표 자료의 예측문제에 있어 Verrall (1990)의 발생연도효과와 경과년도효과만 있는 베이지안 선형모형을 절대연도효과가 있는 모형으로 확장한 모형을 제시하고 이에 대한 추정 방법으로 마르코프 연쇄 몬테칼로 방법을 제안한다. 제안된 모형과 추정 방법은 세 가지 실제 예를 통하여 기존의 방법들에 비해서 일반적으로 작은 상대 예측오차를 제공함을 보였다.

일차 마르코프 잡음 환경에서의 국소 최적 검파 : 2. 성능 분석

이주미,박주호,송익호,오종호,강현구,김선용,Lee, Ju-Mi,Park, Ju-Ho,Song, Iick-Ho,Oh, Jong-Ho,Kang, Hyun-Gu,Kim, Sun-Yong 한국통신학회 2007 韓國通信學會論文誌 Vol.32 No.2C

이 논문의 1부에서는, 곱셈꼴 잡음과 일차 마르코프 덧셈꼴 잡음으로 손상된 약한 신호를 검파할 수 있는 검정 통계량을 얻었다. 이제, 이 2부에서는 이를 바탕으로 여러 검파기의 점근 상대 효율 및 유한 표본 크기 성능을 얻고 견준다. 검파 성능을 처음 뜻한 만큼 내려면 간섭끼리의 의존성을 생각해야 함을 보인다. In Part 1 of this paper, the test statistics are obtained for detecting weak known signals in observations corrupted by multiplicative and first-order Markov additive noise. In this second part, the asymptotic and finite sample-size performance of several detectors are obtained and compared based on the test statistics. From the performance comparisons, it is confirmed that the dependence among interference components need to be taken into account to maintain the detection performance.

일차 마르코프 잡음 환경에서의 국소 최적 검파: 1. 검정 통계량

이주미,박주호,송익호,권형문,김홍직,윤석호,Lee, Ju-Mi,Park, Ju-Ho,Song, Iic-Ho,Kwon, Hyoung-Moon,Kim, Jong-Jik,Yoon, Seok-Ho 한국통신학회 2006 韓國通信學會論文誌 Vol.31 No.10C

이제까지 국소 최적 검파를 다룬 연구들에서는 대부분 관측을 독립이라 두었다. 독립 관측 모형에서 얻은 검파기는 의존성 잡음 성분이 있는 현대 고속 통신 시스템에서 성능이 쾌 떨어질 수 있다. 이 논문의 1부에서는 곱셈 꼴 잡음과 일차 마르코프 덧셈꼴 잡음이 일어나는 환경에서 알려진 약한 신호를 검파할 때 알맞은 검정 통계량을 자세히 살펴본다. 이어, 2부에서는 여러 검파기의 점근 성능과 유한 표본 크기 성능을 얻고 서로 견주어 보며, 성능을 가장 좋게 하려면 간섭끼리의 의존성을 생각하여 검파기를 꾸며야 함을 보인다. In most of the studies on locally optimum detection assumes independent observations. The use of an independent observation model may cause a considerable performance degradation in detection applications of modern high data rate communication systems exhibiting dependence among interference components. In this paper, we address the detection of weak known signals in multiplicative and first order Markov additive noises. In Part 1, the test statistics of the locally optimum detectors are investigated in detail. In Part 2, the asymptotic and finite sample-size performance of several detectors are obtained and compared, confirming that the dependence among interference components need to be taken into account to maintain performance appropriately.

중국인 학습자를 대상으로 한 한국어 종결표현의 지각, 산출 연구 -‘-기만 하다’, ‘-(으)ㄹ 줄 알다’를 중심으로-

이주미 ( Ju Mi Lee ) 국제한국어교육학회 2012 한국어 교육 Vol.23 No.4

When learning the language, prosodic elements are learned firstly. Prosodic elements, however, are learned when learning the foreign language. Rhythm and intonation of the sentence depending on the speaking speed and the language environment. In addition, the meaning is depending on the situation, so exact meaning will be difficult beyond awkward speaking if you do not learn accurate intonation. Therefore, when learning a foreign language, perceive exact sound including allophoneme of the target language and then learn the intonation of pronunciation and sentence according to boundary of Prosodic phrase and Intonational phrase. In this study, intonation pattern was analyzed with the experimental phonetics method according to the meaning about the objects of '-기만 하다' and '-(으). 줄 알다' of Chinese student and Korean speaker. When seeing the results of the experiment, the pitch patterns of (1-1) sentence of 'just look' and (2-1) sentence of 'thought going' are not different between Chinese student and Korean speaker. However, the pitch patterns of the (1-2) sentences of 'just look' and (2-2) sentences of 'thought going' are significantly different between Chinese student and Korean speaker. Even students of intermediate level occur these errors, so pitch education to expression and activity education to perceive the difference are urgently required in the Korean education.

머신러닝을 이용한 부동산 지수 예측 모델 비교

이주미(Lee, Ju-mi),박성훈(Park, Sung-Hoon),조상호(Cho, Sang-ho),김주형(Kim, Ju-Hyung) 대한건축학회 2021 대한건축학회논문집 Vol.37 No.1

As the real estates occupy major portion of domestic households assets, relevant issue has been dealt seriously by the Korean government. However, apartment prices in downtown Seoul, the capital city, have soared despite various policies. Forecasting the real estate market trend has become an important research topic in order to provide information for establishing policies. In the prediction of the real estate market in the previous studies, two research directions were classified as follows: quantitative economic models and machine learning models. Regarding this trend, there was a need for comparative research on machine learning models, emerging methods, that are used to compare and predict various real estate indices. In this study, the machine learning model RF(Random Forest), XGBoost(eXtreme Gradient Boosting), and LSTM (Long Short Term Memory) are used to select suitable machine learning models for selected real estate index and conduct a comparative study to validate predictive power of machine learning models. Apartment sales index, land price index, charter price index, and real estate psychological index using univariate variables are predicted. In addition, RF, XGBoost and LSTM models all tended to be generally marginal with RMSE values of 0.0268, 0.0296, and 0.0259 in charter(Jeonse), Korean traditional pre-deposit rental system, price index data with linear but small variants. This shows that the prediction of the real estate index is deviated from the prediction accuracy of machine learning models depending on the periodic characteristics and data characteristics of the real estate index.

상보적 교수활동이 학문 목적 한국어 학습자의 읽기 능력 활성화에 미치는 효과 연구 -SOLO 분류법을 중심으로-

이주미 ( Lee¸ Ju-mi ) 국제어문학회(구 국제어문학연구회) 2020 국제어문 Vol.0 No.87

본고는 현재 대학에서 수학하는 외국인 학부생들 대상으로 하는 읽기 수업에서 구성주의 이론인 상보적 교수이론을 적용한 수업의 결과로 생산된 학습자들의 요약하기 텍스트를 분석하여 학습자들의 읽기능력이 어떻게 활성화되어 나타나는지 분석해 보는 데 목적이 있다. 이를 위해 학습자들에게 요약하기 과제를 제시하여 학습자들 고유의 읽기 능력으로 텍스트를 어떻게 구조화하였는지 분석하였고, 이후 상호적 교수활동을 한 후 같은 내용의 텍스트를 제시하여 읽기 능력의 구조화에 어떤 변화가 있는지 살폈다. 분석 기준은 SOLO(Structure of the Observed Learning Outcome) 분류법을 이용하였고, 이를 한국어 읽기 능력의 상황에 맞게 단계를 구성하여 분석의 기준으로 삼았다. 분석 결과 학습자들의 읽기 능력은 제시된 텍스트의 내용을 부분적으로만 이해하며, 핵심은 파악하지 못하는 단일 구조화 단계(Uni-structural stage)에서 읽기 맥락을 인식하고 읽은 내용에 대해 지식을 통합하는 관계화 단계 (Relational stage)단계로 발전하였다. 이전에는 시도하지 않았던 지식 변형의 시도가 있었고 문단을 구성하여 요약하려는 시도를 보임으로써 상보적 교수활동이 학습자들의 학업 불안감 해소와 더불어 학습자들의 읽기 능력의 질적인 향상에 긍정적인 영향을 미침을 확인하였다. 한정된 집단에서의 효과를 검증하였다는 점에서 한계가 있지만 향후 학습자 변인이나 교수 환경의 변인에 따라 달리 적용된 연구가 더 활발히 이루어진다면 한국어 읽기 수업의 현장에 새로운 읽기교수법의 지침을 제공해 줄 수 있을 것이다. This study aimed to analyze the aspect of activating learners’ reading abilities by analyzing the learners’ texts for summarizing, produced in the results of the course, applying the reciprocal teaching theory, which is the constructivist theory for reading course targeting the current foreign undergraduates in university. After presenting the task to the learners for summarizing and then analyzing how the texts were structuralized by learners’ inherent reading abilities, the study examined changes in the structuralization of reading abilities by performing the task for summarizing after conducting the reciprocal teaching activities. The Structure of the Observed Learning Outcome (SOLO) Taxonomy was used as the standard of analysis. This was composed of stages suitable for the condition of Korean reading abilities, which was taken as the standard of analysis. The results of the analysis showed that the learners’ reading abilities were developed from the uni-structural stage in which only partial content of the presented texts could be understood without understanding the core to the relational stage in which the learners could perceive the reading context and integrate the knowledge about the content already read. In addition to the attempt to transform knowledge, which had not been attempted before, there was also an attempt to summarize by composing sentences. The results of this study verified that Reciprocal Teaching would have positive effects on the relief of learners’ academic anxiety and the qualitative improvement of their reading abilities. Despite the limitation in the aspect of verifying the effects on a limited group, more active studies in the future applying different learner variables or teaching environment variables could provide guidelines for new teaching methods of reading in the Korean reading course.

중국인 한국어 학습자의 숙달도에 따른 유의어 사용 양상 분석 연구 : 한국어와 중국어의 언어 대조를 중심으로

이주미(Lee, ju mi),이연정(Lee, yeon jeong) 학습자중심교과교육학회 2021 학습자중심교과교육연구 Vol.21 No.6

본고는 유의어 ‘이용하다’와 ‘사용하다’가 한국어와 중국어 두 언어에서 어떤 어휘들과 공기관계를 갖는지 밝히고 이러한 차이가 한국어 학습자의 유의어 사용 양상에 미치는 영향을 분석해 보는 데 목적이 있다. 이를 위해 먼저 사전적 어휘 정보를 통해 한국어와 중국어에서 ‘이용하다’와 ‘사용하다’가 가지는 의미와 기능을 분석하여 두 동사의 의미 범주의 차이를 밝혔다. 이후 설문 조사를 통해 중국인 학습자들의 ‘이용하다’와 ‘사용하다’ 사용에 대한 양상을 조사하였고 이를 숙달도에 따라 분석하였다. 분석결과 중급학습자들은 ‘이용하다’와 ‘사용하다’의 혼용이 많아 두 어휘의 의미범주에 대해 명확히 인지하지 못하는 것으로 나타났다. 이는 한국어에 비해 중국어에서 ‘이용하다’와 ‘사용하다’가 취하는 목적어의 범위가 넓고 유사하기 때문인 것으로 분석되었다. 중급학습자들에 비해 고급학습자들이 유의어의 의미 범주를 명확히 구분하고 있었으나 여전히 모국어의 영향을 받아 혼용되는 경우가 나타났다. 이에 중급 단계 이상의 숙달도를 가진 한국어 학습자들에게는 두 언어에서 어휘의 의미가 어떻게 나타나는지 명확히 제시해야 할 필요성이 있음을 제안하였다. 특히 용법을 중심으로 유의어에 대한 변별 정보를 명시적(explicit)으로 제시하여 학습자들이 스스로 목표어에 대한 의미범주를 스스로 인지할 수 있도록 교육해야 한다. The purpose of this research is to investigate what co-occurrence relation the synonym ‘utilize’ and ‘use’ has with other vocabulary in Korean and Chinese and analyze the influence such difference causes on the aspect of using the synonym by Korean learners. For that, the difference of meaning between two verbs was clarified by analyzing the meaning and function of ‘utilize’ and ‘use’ have in Korean and Chinese through dictionary vocabulary information. After that, the aspect of using ‘utilize’ and ‘use’ by Chinese learners was investigated through questionnaire survey and the proficiency was analyzed. As the results of analysis, it showed that the learners in middle class could not recognize the meaning of two words clearly as they chose ‘utilize’ and ‘use’ in mixed way. It was analyzed that is because the range of objects that ‘utilize’ and ‘use’ take in Chinese was wider and similar comparing to Korean. Though the learners in high class distinguish the meaning of the synonym correctly, some cases showed they were still influenced by their mother tongue. So, the necessity of defining how such two words mean to the learners with the proficiency higher than middle class was suggested clearly. The distinctive information about the synonym should be suggested explicitly focusing on the usage especially for education so the learners can recognize the meaning of the objectives by themselves.

자발성을 통해 본 신경숙 소설의 주인공 성격 연구 -『외딴방』을 중심으로-

이주미 ( Ju Mi Lee ) 한국현대문예비평학회 2009 한국문예비평연구 Vol.- No.28

대학 수학 한국어 학습자의 글쓰기 교육에 대한 내용 범위 연구 -요구분석에 따른 우선순위결정을 중심으로-

이주미 ( Ju Mi Lee ) 국제한국어교육학회 2016 한국어 교육 Vol.27 No.2

The purpose of this study is to consider the scope and order of priority in writing education by conducting a needs analysis targeted for learners studying in college at present. The teaching and learning context of writing education in college is decided depending on the purpose of the college or the judgment of professors rather than an accurate judgment of learner factors. To improve the writing skills of learners, it is crucial to accurately evaluate the present level of learners. Therefore, this study conducted a needs survey by using the questionnaire technique targeting learners taking a writing class as a liberal arts course in college at present. The survey is formed to analyze two measures, such as the level of performance and the level of motivation at the same time. The T-test is used to decide the order of priority, but the weakness of the t-test is supplemented by using a Borich coefficient. In addition, by using an IPA(Importance-Performance Analysis) Matrix, the order of priority regarding the difference between two levels was suggested; thus, raising concerns of the validity of the study. If the level of performance and the level of motivation of learners studying in college are accurately understood, guidelines can be suggested to sort and rank effective instruction for learners, which can contribute greatly to the selection of tasks and activities, and the development and evaluation of teaching materials. (Korea University)

자기결정성 이론에 기반한 한국어 학습자의 학문적 학습 동기 항목의 척도 개발 및 타당화 연구

이주미 ( Lee Ju-mi ) 이중언어학회 2021 이중언어학 Vol.85 No.-

The purpose of this paper is to develop items to measure the academic learning motivation of Korean learners studying at university and to verify the validity of them based on the self-determination theory, To this end, we analyzed AMTB, SRQ-A, SMS, and KASRQ, which are learning motivation measurement tools developed based on the self-determination theory, and extracted measurement items by considering the learning situational factors of foreign Korean learners. The extracted items were verified for validity according to the consensus of the expert panel. In most cases, the validity process was carried out in 2 rounds, but the external regulation item had a low alpha value in the reliability analysis of the 2nd result, so 3rd verification was additionally performed through an interview with the expert panel. Each item validated in this way was classified into items with a Cronbach's-α value of 0.7 or higher through reliability analysis, and as a result, 56 items were selected. Cronbach's-α for the 56 items classified with measurement tools was 0.887, indicating an excellent level of reliability and high internal consistency as measurement tools. (Ajou University)