• The Journal of Society for e-Business Studies
• /
• v.14 no.4
• /
• pp.47-57
• /
• 2009

Collaborative tagging systems allow users to attach tags to diverse sharable contents in social networks. These tags provide usefulness in reusing the contents for all community members as well as their creators. Three-dimensional data composed of users, items, and tags are used in the collaborative tag-based recommendation. They are generally more voluminous and sparse than two-dimensional data composed of users and items. Therefore, there are many difficulties in applying existing collaborative filtering methods directly to them. Latent factor models, which are also successful in the area of collaborative filtering recently, discover latent features(factors) for explaining observed values and solve problems based on the features. However, establishing the models require much time and efforts. In order to apply the latent factor models to three-dimensional collaborative filtering data, we have to overcome the difficulty of establishing them. This paper proposes various methods for determining preferences of users to items via establishing an intuitive model by assuming tags used for items as latent factors to users and items respectively. They are compared using real data for concluding desirable directions.

• Journal of the Society of Disaster Information
• /
• v.11 no.1
• /
• pp.23-34
• /
• 2015

This Study analyzed heavy snow properties according to the area that was based by winter weather properties and the damage data by the heavy snow among each local government of Chungcheongbuk-do. The result of analysis, Jecheon-si and Boeun-gun are represented the highest dangerous regions by potential degree of risk by average amount of snowfall for 35 years. But, the potential degree of risk by maximum amount of snowfall for 35 years is different with it. Cheongju-si and Youngdong-gun, Goesan-gun, Boeun-gun are represented the highest dangerous regions. Examining the frequency of regions with potential danger factors according to the characteristics of heavy snowfall, Boeun-gun and Jecheon-si, Goesan-gun, Youngdong-gun, Cheongju-si is derived the highest dangerous regions in Chungcheongbuk-do.

• Journal of the Society of Disaster Information
• /
• v.17 no.1
• /
• pp.68-80
• /
• 2021

Purpose: In this study, potential demographic, social, and economic factors causing tuberculosis were identified, and eight cities in South Chungcheong Province were compared and analyzed with the IPA method. Method: The factors potentially affecting the prevalence of tuberculosis were categorized demographically, socially, and economically, and selected through brainstorming. Furthermore, potential factors affecting tuberculosis were derived using the revised IPA. Based on this analysis, areas with a potential risk of tuberculosis were classified, and the following policy implications were suggested. Result: The analysis found the three cities of Nonsan, Boryeong, and Gongju to have the highest potential risk of tuberculosis, and the frequency of potential risk factors in the above three cities to be 6 or more. Thus, an urgent policy response to prevent tuberculosis in these regions is required. Conclusion: According to the results of this study, it is necessary to take potential risk factors into account when promoting tuberculosis prevention policies and projects in South Chungcheong Province.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2015.07a
• /
• pp.335-337
• /
• 2015

사고 준사고 선박의 경우 하인리히 법칙에 의거 사고 전 잠재적 요인을 내재하고 있는 선박이 그 예후를 미리 보이는 경우가 많다. 이러한 잠재적 요인의 집계 및 통계를 통해 사고 유발 가능선박을 분리하고, 이 선박들에 한해서는 기존의 사고 후 대책 마련 방식에서 벗어나 대형 사고를 예방하기 위한 방식으로 접근하는 것이 철저하게 필요하다고 본다. 그리하여 동해 VTS 관제 구역 내 사고 유발 가능 선박에 나타났던 잠재적 요인의 예시들과, 기존 접근 방법의 한계점에 대해 논하고, 분리된 선박에 대해 어떠한 식으로 관리할 것인지 논해보고자 한다.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.249-263
• /
• 2023

Collaborative Filtering, a representative recommendation system methodology, consists of two approaches: neighbor methods and latent factor models. Among these, the latent factor model using matrix factorization decomposes the user-item interaction matrix into two lower-dimensional rectangular matrices, predicting the item's rating through the product of these matrices. Due to the factor vectors inferred from rating patterns capturing user and item characteristics, this method is superior in scalability, accuracy, and flexibility compared to neighbor-based methods. However, it has a fundamental drawback: the need to reflect the diversity of preferences of different individuals for items with no ratings. This limitation leads to repetitive and inaccurate recommendations. The Adaptive Deep Latent Factor Model (ADLFM) was developed to address this issue. This model adaptively learns the preferences for each item by using the item description, which provides a detailed summary and explanation of the item. ADLFM takes in item description as input, calculates latent vectors of the user and item, and presents a method that can reflect personal diversity using an attention score. However, due to the requirement of a dataset that includes item descriptions, the domain that can apply ADLFM is limited, resulting in generalization limitations. This study proposes a Generalized Adaptive Deep Latent Factor Recommendation Model, G-ADLFRM, to improve the limitations of ADLFM. Firstly, we use item ID, commonly used in recommendation systems, as input instead of the item description. Additionally, we apply improved deep learning model structures such as Self-Attention, Multi-head Attention, and Multi-Conv1D. We conducted experiments on various datasets with input and model structure changes. The results showed that when only the input was changed, MAE increased slightly compared to ADLFM due to accompanying information loss, resulting in decreased recommendation performance. However, the average learning speed per epoch significantly improved as the amount of information to be processed decreased. When both the input and the model structure were changed, the best-performing Multi-Conv1d structure showed similar performance to ADLFM, sufficiently counteracting the information loss caused by the input change. We conclude that G-ADLFRM is a new, lightweight, and generalizable model that maintains the performance of the existing ADLFM while enabling fast learning and inference.

• 한국항공운항학회:학술대회논문집
• /
• 2016.05a
• /
• pp.79-83
• /
• 2016

1990년부터 14년간 총 65건의 회전익 사고가 발생하여 연 평균 4.6건의 사고가 발생하였다. 항공대국으로 안전한 수준에 이르기 위하여 회전익 항공기 사고의 잠재적 조건 선행 분석자료 및 연구결과 등을 활용하여 회전익 항공기 항공안전대책을 살펴보았다. 소방항공의 경우 인적요인이 사고 발생원인의 57%로 주 요인으로 나타났으며, 국가기관에서도 2000년 이후 사고 중에서 인적요인이 주 요인으로 인적요인이 중요한 개선대상으로 밝혀졌다. 항공안전 전문가들과 인터뷰를 실시하여 회전익 항공기의 잠재 조건에 대해 논의하였다.

• Journal of Korean Society of Transportation
• /
• v.28 no.4
• /
• pp.119-128
• /
• 2010

Mode choice behavior is associated with travelers' latent behavior that is an unobservable preference to travel behavior or mode characteristics. This paper specifically addresses the problem of unobservable factors, that is latent behavior, in mode choice models. Consideration of latent behavior in mode choice models reduces the errors that come from unobservable factors. In this study, the authors defined the latent variables that mean a quantitative latent behavior factors, and developed the combined RP/SP model with latent variables using the mode choice behavior survey data. The data has traveler's revealed preference of existent modes along the Han River and stated preference of new water transit on the Han River. Also, The data has travelers' latent behavior. Latent variables were defined by factor analysis using the latent behaviour data. In conclusion, it is significant that the relationship between traveler's latent behavior and mode choice behavior. In addition, the goodness-of-fit of the mode choice models with latent variables are better than the model without latent variables.

• 한국HCI학회:학술대회논문집
• /
• 2007.02b
• /
• pp.685-692
• /
• 2007

본 연구에서는 로저스의 개혁확산이론(Rogers, 2003)과 데이비스의 기술수용모형(Davis, 1989)을 바탕으로 실감공간기술에 대한 잠재적 사용자의 태도를 분석하였다. 개혁확산이론과 기술수용모형을 바탕으로 한 선행 연구들을 통해 새로운 미디어의 채택에 영향을 미치는 다양한 요인들을 고찰할 수 있었다. 개혁확신이론을 통해서는 개인의 심리적 수준은 물론 인구통계학적 수준, 사회적 수준 등 다양한 요인과 그 하부 요인이 도출되었으며, 기술수용모형을 통해서는 개혁확산이론을 통해 도출된 다양한 변인들이 실감공간기술에 적용될 수 있다는 가설을 설계할 수 있었다. 개혁확산 이론과 기술수용모형을 통해 이미 설명된 기존의 뉴미디어들과 달리 실감공간기술이 개발 진행 중이라는 점과 그 다양한 발전 가능성이라는 특징을 고려했을 때, 실감공간기술의 확산을 어떻게 예측하고 설명할 수 있는지 고찰하는 것 또한 본 연구의 목적이라고 할 수 있다. 본 연구에서는 실감공간기술에 대한 잠재적 사용자들의 사용의도에 영향을 미치는 요인의 영향력을 살펴봄으로써 실감공간기술의 다양한 활용 방안에 대하여 모색하였다.

• Journal of Korean Society of Transportation
• /
• v.28 no.3
• /
• pp.99-107
• /
• 2010

Analyzing mode choice among transportation demand estimate procedures is complicated and understanding characteristics of travelers is also difficult. Generally, it is well known that traveler choose mode considering psychometric factors and characteristic besides socio-demographic indicators. Accordingly, many researches has investigated on methodology that can be applied in mode choice to reflect psychometric factor or specific preference. Latent Class Analysis among various studies is recognized as the theoretically potential approach. This study focuses on class segmented using latent class cluster to analyze impact that included psychometric factors and characteristics on mode choice. It also provides evidence that mode choice model for each class and mode choice model not considering latent class are different. This study based on citizen's stated preference and revealed preference on a new transit on the Han river shows that latent class cluster analysis is the potential approach considering latent preference.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.3
• /
• pp.71-81
• /
• 2019

To extend the potential benefits of error, the current study examined factors that affect students' error perception in the classroom. An experimental design was used to measure relations of classroom goal structure, feedback, and social relationships on students' perception of error. A total of 316 fourth-, fifth-, and sixth-grade elementary students participated as part of their regular class curriculum. Self-reported questionnaires were administered to measure students' perception of errors and relationships with teacher and peers, and then students were manipulated by classroom goal structure and feedback. Multiple regression analysis results suggested that students' perception of learning from error was affected mostly by relationships with peers, followed by relationships with teacher and the type of feedback. Students' perception of risk taking for error was also affected by relationships with peers and teacher, followed by the classroom goal structure. However, classroom goal structure and feedback did not affect their perception of thinking about error to improve their learning as well as error strain. These results imply how the classroom climate should be structured to improve perception of errors to improve student's learning.