• 말소리와 음성과학
• /
• 제15권3호
• /
• pp.83-88
• /
• 2023

This study investigates the fine-tuning of large-scale Automatic Speech Recognition (ASR) models, specifically OpenAI's Whisper model, for domain-specific applications using the KsponSpeech dataset. The primary research questions address the effectiveness of targeted lexical item emphasis during fine-tuning, its impact on domain-specific performance, and whether the fine-tuned model can maintain generalization capabilities across different languages and environments. Experiments were conducted using two fine-tuning datasets: Set A, a small subset emphasizing specific lexical items, and Set B, consisting of the entire KsponSpeech dataset. Results showed that fine-tuning with targeted lexical items increased recognition accuracy and improved domain-specific performance, with generalization capabilities maintained when fine-tuned with a smaller dataset. For noisier environments, a trade-off between specificity and generalization capabilities was observed. This study highlights the potential of fine-tuning using minimal domain-specific data to achieve satisfactory results, emphasizing the importance of balancing specialization and generalization for ASR models. Future research could explore different fine-tuning strategies and novel technologies such as prompting to further enhance large-scale ASR models' domain-specific performance.

• 한국정보시스템학회지:정보시스템연구
• /
• 제26권4호
• /
• pp.17-37
• /
• 2017

Purpose Customer Loyalty is the most important factor of customer relationship management (CRM). Especially in retailing industry, where customers have many options of where to spend their money. Classifying loyal customers through customers' data can help retailing companies build more efficient marketing strategies and gain competitive advantages. This study aims to construct classification models of distinguishing the loyal customers within a Korean retailing company using data mining techniques with R language. Design/methodology/approach In order to classify retailing customers, we used combination of support vector machines (SVMs) and other classification algorithms of machine learning (ML) with the support of recursive feature elimination (RFE). In particular, we first clean the dataset to remove outlier and impute the missing value. Then we used a RFE framework for electing most significant predictors. Finally, we construct models with classification algorithms, tune the best parameters and compare the performances among them. Findings The results reveal that ML classification techniques can work well with CRM data in Korean retailing industry. Moreover, customer loyalty is impacted by not only unique factor such as net promoter score but also other purchase habits such as expensive goods preferring or multi-branch visiting and so on. We also prove that with retailing customer's dataset the model constructed by SVMs algorithm has given better performance than others. We expect that the models in this study can be used by other retailing companies to classify their customers, then they can focus on giving services to these potential vip group. We also hope that the results of this ML algorithm using R language could be useful to other researchers for selecting appropriate ML algorithms.

• 대한간호학회지
• /
• 제42권1호
• /
• pp.36-47
• /
• 2012

Purpose: This study was to explore adaptation experience to family among women who immigrated for marriage. Specific aims were to identify problems immigrant women face as family members and how they interact with other family members. Methods: Grounded theory methodology was utilized. Data were collected from iterative fieldwork with individual in-depth interviews from 6 immigrant women as key informants, and 2 of their husbands and 2 of their mothers-in-law as general informants. Results: Through constant comparative analysis, a core category emerged as "tearing down the wall in communicating". Causal conditions were feeling frustrated in one's expectations, differences in language and life style, differences in recognition, and perceptions of discrimination and prejudice. Strategies were learning the Korean language, learning Korean culture, managing stress, mediating differences between family members, and introspecting. Intervening factors were support systems, burdens of child-rearing, and the condition of one's health. Consequences were rooting oneself in one's family and accepting one's life as it is. Conclusion: Results of the study indicate that there is a need for nurses to understand differences in communication with family members among immigrant women and to provide information and emotional support to improve the adaptation of these women to their Korean families.

• 정보교육학회논문지
• /
• 제12권3호
• /
• pp.323-332
• /
• 2008

본 연구에서는 초등학생의 프로그래밍 학습을 효과적으로 조력하기 위해 새로운 교육용 프로그래밍 언어인 스크래치를 활용한 프로그래밍 학습의 가능성을 제시하고자 하였다. 스크래치 프로그래밍 학습 내용은 프로그래밍 과정에서의 학습자의 내적 동기 유발을 위한 전략과 복잡한 인지 능력 향상을 위한 창의적 문제해결 수업모형(CPS)을 토대로 구성하였다. 설계된 학습 내용을 초등학교 6학년 재량활동 시간에 적용한 결과, 스크래치 프로그래밍 학습은 학습자의 내재적 동기와 문제해결력 향상에 효과가 있는 것으로 나타났다. 최근 정보통신기술교육 운영지침의 개정으로 프로그래밍 교육과 알고리즘 교육이 초등학교에서도 필요하게 되었고 학교 현장의 여건 등을 고려해보면 본 연구를 통해 설계된 교수 학습 전략을 기반으로 한 스크래치 프로그래밍 학습은 의미 있는 선택이며 기존 프로그래밍 교육의 문제점 해소를 위한 적절한 대안이 될 수 있을 것이다.

• 영어어문교육
• /
• 제17권2호
• /
• pp.199-218
• /
• 2011

This study is to explore one of efficient procedures in producing English textbooks for secondary school students. According to a series of changes in the National Curriculum, new textbooks have been selected and used in English classes. Textbooks are one of the fundamental factors in teaching and learning languages together with learners and teachers. So this study emphasizes the significance of textbooks and presents the practical model of producing English textbooks including activity books, from major aspects such as planning, writing, editing, selecting, etc. The current government has made continuing efforts to improve English education development by administrating innovative policies [strategies]. However, there still remain lots of difficulties in this gigantic task, which is not an exception in the matter of textbooks. Therefore, to provide students with better textbooks, the government should not only invest great funds but also renovate the present polluted system of selecting textbooks.

• Journal of Multimedia Information System
• /
• 제6권4호
• /
• pp.173-178
• /
• 2019

Mongolian historical and cultural monuments on settlement areas of stone inscriptions, stone images, rock-drawings, remains of cities, architecture are still telling us their stories. These monuments depict the understanding of the word, philosophical and artistic outlook, beliefs, religion, national art, language, culture and traditions of Mongols [1]. Nowadays computer science, especially computer vision is applying in the other science fields. The main problem is how to apply and which algorithm can detect and classify the objects correctly. In this paper, we propose a method to detect object from Mongolian nomadic environment images. This work proposes a method for object detection that is the combination of the binary operations in the edge detection results. We found out the best method and parameters of state-of-the-art machine learning algorithms. In experimental result, we evaluate our results with 10-fold cross validation and split 66% strategies.

• 영어어문교육
• /
• 제17권1호
• /
• pp.101-124
• /
• 2011

This paper investigated the relationships among Korean EFL learners' listening anxiety, listening strategy use, and listening proficiency. One hundred and forty four Korean college students who were enrolled in the required practical English classes participated in this study. Questionnaires related to students' listening strategy use and listening anxiety were administered and a TOEIC listening comprehension test was given to measure the students' listening proficiency. The one-way ANOVA was used to analyze the data. The findings of this study are that the students' listening performance is positively correlated with their strategy use and negatively correlated with their anxiety level, and their strategy use is negatively correlated with their anxiety level. The results suggest that successful learning will occur when anxiety is reduced and when the use of strategies is encouraged more often. The pedagogical implications for EFL educators and teachers are described.

• 영어어문교육
• /
• 제15권2호
• /
• pp.1-25
• /
• 2009

The purpose of this study is to explore the potential of peer coaching for EFL teachers' professional development. For this study, 12 college teachers in Korea participated in a 10-week program. They were 7 males and 5 females, ranging in age from 24 to 37 years. Data were collected through semi-structured interviews. Reflective analysis was used to analyze individual interview data. From the findings, two significant categories of peer coaching were identified: positive and negative responses to peer coaching experience. However, the overriding themes that emerged from the data were the benefits of peer coaching. The participants were almost unanimous in their acknowledgement of the advantages of peer coaching, such as reflective support through other's eyes, improved working environments, greater teaching strategies, higher professional self-esteem, and awareness of self-directed learning. Negative responses also appeared, mostly in regard to the working principles of implementation; the major issues of difficulties were time management, complexities of implementation procedure, stress and personal vulnerability, and relative lack of reflection and feedback skills. Demonstrating the participants' experiences towards the peer coaching program, this study provides EFL teachers with useful insights into peer coaching as an effective tool of their professional development.

• 영어어문교육
• /
• 제15권2호
• /
• pp.27-49
• /
• 2009

This study aimed to explore EFL learners' beliefs and preferences about reading tasks and to examine the reading processes that they use for making sense of text. The subjects were comprised of 107 college students who were non-English majors and aged 19-28 years. Based on scores achieved on a reading comprehension test, they were divided into two groups (more-skilled and less-skilled readers) and asked to respond to a survey in class. The results of the survey revealed that: (1) a majority rate themselves as fair readers, which might be indicative of the insecurity they feel toward L2 reading; (2) authentic texts (especially magazines) and popular media appear to be their favorite reading materials; (3) unknown vocabulary is a major impediment to their L2 reading comprehension; (4) the more-skilled readers manifest a meaning centered view of reading, whereas the less-skilled readers center on vocabulary; and (5) both groups employ a multistrategic approach to L2 reading; however, the less-skilled readers are less successful in determining the meaning of unknown vocabulary. Pedagogical implications for EFL classroom teachers are provided.

• International Journal of Computer Science & Network Security
• /
• 제22권5호
• /
• pp.135-142
• /
• 2022

Despite global medical advancements, many patients are misdiagnosed, and more people are dying as a result. We must now develop techniques that provide the most accurate diagnosis of heart disease based on recorded data. To help immediate and accurate diagnose of heart disease, several data mining methods are accustomed to anticipating the disease. A large amount of clinical information offered data mining strategies to uncover the hidden pattern. This paper presents, comparison between different classification techniques, we applied on the same dataset to see what is the best. In the end, we found that the Random Forest algorithm had the best results.