http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
토픽모델링을 활용한 상담과정에서 상담자와 내담자 간 언어변화
김하선(Ha-Seon Kim),조남옥(Nam-Ok Cho),이윤주(Yoon-Joo Lee) 학습자중심교과교육학회 2023 학습자중심교과교육연구 Vol.23 No.20
목적 본 연구의 목적은 상담과정에서 이루어진 상담자와 내담자 간의 상담 내용을 토픽모델링을 활용하여 주요 주제를 추출하며 시간의 흐름에 따른 언어변화를 탐색하는 데 있다. 방법 한 상담사가 주 1회 50분 동안 총 5회에 걸쳐 내담자 A와 내담자 B에게 개별상담을 진행한 내용과 추수상담 1회를 포함한 상담 축어록을 전사한 후, 정제과정을 거친 뒤 빈도분석과 토픽모델링을 실시하였다. 결과 빈도 분석 결과, 내담자 A의 상담에서 주요 핵심어는 ‘생각’, ‘사람’, ‘마음’ 등이 나타났으며, 내담자 B의 상담에서는 ‘엄마’, ‘아이’, ‘이야기’ 등이 주로 등장하였다. LDA토픽모델링 결과, 내담자 A의 주요 토픽은 ‘생각과 기분’, ‘자신의 생활’ 등으로, 상담자의 토픽은 ‘남편과 마음’, ‘사람의 마음’ 등 5개씩 추출되었다. 내담자 B의 주요 토픽은 ‘교수 생각’, ‘칭찬 이야기’ 등으로, 상담자의 토픽은 ‘걱정된 이야기’, ‘엄마 공부’ 등 8개씩 추출되었다. 또한, DTM토픽모델링 결과, 내담자 A와 B 모두 상담 초기와 종결 회기에서의 토픽 출현이 높았으며, 상담과정 중에는 낮았다. 반면, 상담자는 상담 초기부터 종결 회기까지 토픽의 출현이 높았으며, 추수상담에서는 내담자와 상담자 모두 토픽 출현이 낮게 나타났다. 결론 본 연구를 통해, 상담과정 중 핵심 주제어와 토픽을 추출하여 내담자의 관심사와 주제를 파악하고 토픽 변화를 분석함으로써, 상담자와 내담자 간 상호작용에서 특정 주제와 언어가 어떻게 변화되는지 탐색하는 데 도움이 되었다. 이러한 연구 결과는 상담자가 내담자의 주요 관심사에 초점을 맞춘 개인상담을 진행하고 있는지 확인하여 상담과정의 질을 개선하는 데 중요한 정보를 제공함을 시사한다. 뿐만 아니라, 개인상담의 내용을 분석하는 데 토픽모델링이라는 인공지능 분석기법을 활용하여, 상담과정에서 상담자와 내담자 간 언어변화를 제시하는 데 의의를 두고 있다. Objectives The aim of this study is to utilize counseling content between counselors and clients during the coun-seling process using topic modeling, extract key themes, and investigate language changes over time. Methods A single counselor conducted individual counseling sessions with Client A and Client B, each lasting 50 minutes, once a week for a total of 5 sessions. The transcripts of these counseling sessions, including a follow-up session, underwent refinement before undergoing frequency analysis and topic modeling. Results In the frequency analysis, key keywords in Client A's counseling sessions included ‘thoughts,’ ‘people,’ and ‘emotions,’ while in Client B's counseling sessions, ‘mother,’ ‘child,’ and ‘stories’ predominated. LDA topic modeling revealed that Client A's major topics were ‘thoughts and emotions’ and ‘personal life,’ while topics such as ‘husband and emotions’ and ‘people's emotions’ were extracted for the counselor, each occurring 5 times. For Client B, primary topics included ‘professor's thoughts’ and ‘complimentary stories,’ while for the counselor, top-ics like ‘worrisome stories’ and ‘mother's studies’ were extracted, each occurring 8 times. Additionally, DTM topic modeling results showed that both Client A and B had higher topic appearances at the beginning and end stages of counseling, with lower appearances during the counseling process. In contrast, the counselor had a high topic appearance from the beginning to the end, while in the follow-up session, both the client and the counselor showed lower topic appearances. Conclusions Through this study, extracting core keywords and topics during the counseling process allowed for the identification of client interests and themes, as well as the analysis of topic changes over time. This in-formation can assist counselors in focusing on the client's primary concerns, thus improving the quality of the counseling process. Furthermore, the use of artificial intelligence analysis techniques such as topic modeling in analyzing the content of individual counseling sessions is significant in presenting language changes between counselors and clients during the counseling process.
노이즈 레벨 및 유사도 평가 기반 저선량 조건의 전산화 단층 검사 영상에서의 비지역적 평균 알고리즘의 최적화
정하선(Ha-Seon Jeong),김이준(Ie-Jun Kim),박수빈(Su-Bin Park),박수연(Suyeon Park),오윤지(Yunji Oh),이우석(Woo-Seok Lee),서강현(Kang-Hyeon Seo),이영진(Youngjin Lee) 대한방사선과학회(구 대한방사선기술학회) 2024 방사선기술과학 Vol.47 No.1
In this study, we optimized the FNLM algorithm through a simulation study and applied it to a phantom scanned by low-dose CT to evaluate whether the FNLM algorithm can be used to obtain improved image quality images. We optimized the FNLM algorithm with MASH phantom and FASH phantom, which the algorithm was applied with MATLAB, increasing the smoothing factor from 0.01 to 0.05 with increments of 0.001 and measuring COV, RMSE, and PSNR values of the phantoms. For both phantom, COV and RMSE decreased, and PSNR increased as the smoothing factor increased. Based on the above results, we optimized a smoothing factor value of 0.043 for the FNLM algorithm. Then we applied the optimized FNLM algorithm to low dose lung CT and lung CT under normal conditions. In both images, the COV decreased by 55.33 times and 5.08 times respectively, and we confirmed that the quality of the image of low dose CT applying the optimized FNLM algorithm was 5.08 times better than the image of lung CT under normal conditions. In conclusion, we found that the smoothing factor of 0.043 among the factors of the FNLM algorithm showed the best results and validated the performance by reducing the noise in the low-quality CT images due to low dose with the optimized FNLM algorithm.
인공지능 기반 유사한 형상을 가진 표적 식별 가능성 확인을 위한 레이더 데이터 분석
김아란(A Ran Kim),김하선(Ha Seon Kim),강창호(Chang Ho Kang),김선영(Sun Young Kim) 제어로봇시스템학회 2022 제어·로봇·시스템학회 논문지 Vol.28 No.4
In this work, we analyzed radar data to check the feasibility of identifying targets with similar shapes based on artificial intelligence. Among radar measurements, radar cross section (RCS) and high-resolution range profile (HRRP) were selected and used as the classification metrics. Before performing artificial intelligence learning, the structural similarity index measure was selected as the performance index and used to verify the feasibility of target classification. We modeled various targets with similar shapes and then obtained radar data using Ansys HFSS. From similar test results, we confirmed that targets with similar shapes could be identified and the possibility of classification in the case of HRRP is higher than that in the case of RCS.