• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.8
• /
• pp.2587-2605
• /
• 2022

In order to efficiently detect community structure in complex networks, community detection algorithms can be designed from the perspective of node similarity. However, the appropriate parameters should be chosen to achieve community division, furthermore, these existing algorithms based on the similarity of common neighbors have low discrimination between node pairs. To solve the above problems, a noval community detection algorithm using closeness similarity based on common neighbor node clustering entropy is proposed, shorted as CSCDA. Firstly, to improve detection accuracy, common neighbors and clustering coefficient are combined in the form of entropy, then a new closeness similarity measure is proposed. Through the designed similarity measure, the closeness similar node set of each node can be further accurately identified. Secondly, to reduce the randomness of the community detection result, based on the closeness similar node set, the node leadership is used to determine the most closeness similar first-order neighbor node for merging to create the initial communities. Thirdly, for the difficult problem of parameter selection in existing algorithms, the merging of two levels is used to iteratively detect the final communities with the idea of modularity optimization. Finally, experiments show that the normalized mutual information values are increased by an average of 8.06% and 5.94% on two scales of synthetic networks and real-world networks with real communities, and modularity is increased by an average of 0.80% on the real-world networks without real communities.

• Journal of Software Engineering Society
• /
• v.25 no.4
• /
• pp.69-74
• /
• 2012

This article suggests applicability to community detection algorithm from module recovering process of software architecture through compare to software clustering metric and community dectection metric. in addition to, analyze mutual relation and difference between separated module and measurement value of typical clustering algorithms and community detection algorithms. and then only sugeested several kinds basis that community detection algorithm can use to recovering module view of software architecture and, by so comparing measurement value of existing clustering metric and community algorithms, this article suggested correlation of two result data.

• Journal of Korea Association of Health Promotion
• /
• v.3 no.2
• /
• pp.137-146
• /
• 2005

The Community based cancer screening program passed in 1960 was a milestone for initiating a national and local health program in Japan. And since then local governments and Cancer Society have been developing and providing cancer screening programs of Stomach, Cervix, Breast and Colorectum for population. To apply the effectiveness of community based cancer screening program, it is important to understand the key issue related to cancer screening participation of population and technology of cancer detection. The purpose of this study was to understand the community based cancer screening program in Japan, and to apply the information for establishment of community based cancer screening program in Korea. The characteristics of community based cancer screening program in Japan were as follows. The first, community based cancer screening program was implemented by the National Health and Medical Services Law for the Aged since 1983. The second, Cancer Society and Cancer Detection Center were core for cancer screening program. The third, the budget for cancer screening program was established by the National Health and Hygiene. The fourth, the continuous quality control for medical staff was provided by Cancer Society and Cancer Detection Center The fifth, the efforts for the promotion of cancer screening rate.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.5
• /
• pp.2142-2161
• /
• 2018

Distance dynamics model is an excellent model for uncovering the community structure of a complex network. However, the model has poor robustness. To improve the robustness, we design an enhanced distance dynamics model based on Ego-Leader and propose a corresponding community detection algorithm, called E-Attractor. The main contributions of E-Attractor are as follows. First, to get rid of sensitive parameter ${\lambda}$, Ego-Leader is introduced into the distance dynamics model to determine the influence of an exclusive neighbor on the distance. Second, based on top-k Ego-Leader, we design an enhanced distance dynamics model. In contrast to the traditional model, enhanced model has better robustness for all networks. Extensive experiments show that E-Attractor has good performance relative to several state-of-the-art algorithms.

• Journal of Internet Computing and Services
• /
• v.24 no.6
• /
• pp.1-11
• /
• 2023

Federated learning (FL) is a ground breaking machine learning paradigm that allow smultiple participants to collaboratively train models in a cloud environment, all while maintaining the privacy of their raw data. This approach is in valuable in applications involving sensitive or geographically distributed data. However, one of the challenges in FL is dealing with heterogeneous and non-independent and identically distributed (non-IID) data across participants, which can result in suboptimal model performance compared to traditionalmachine learning methods. To tackle this, we introduce FedGCD, a novel FL algorithm that employs Graph Neural Network (GNN)-based community detection to enhance model convergence in federated settings. In our experiments, FedGCD consistently outperformed existing FL algorithms in various scenarios: for instance, in a non-IID environment, it achieved an accuracy of 0.9113, a precision of 0.8798,and an F1-Score of 0.8972. In a semi-IID setting, it demonstrated the highest accuracy at 0.9315 and an impressive F1-Score of 0.9312. We also introduce a new metric, nonIIDness, to quantitatively measure the degree of data heterogeneity. Our results indicate that FedGCD not only addresses the challenges of data heterogeneity and non-IIDness but also sets new benchmarks for FL algorithms. The community detection approach adopted in FedGCD has broader implications, suggesting that it could be adapted for other distributed machine learning scenarios, thereby improving model performance and convergence across a range of applications.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.273-276
• /
• 2019

There is a lot of information in our world, quick access to the most accurate information or finding the information we need is more difficult and complicated. The recommendation system has become important for users to quickly find the product according to user's preference. A social recommendation system using community detection based on label propagation is proposed. In this paper, we applied community detection based on label propagation and collaborative filtering in the movie recommendation system. We implement with MovieLens dataset, the users will be clustering to the community by using label propagation algorithm, Our proposed algorithm will be recommended movie with finding the most similar community to the new user according to the personal propensity of users. Mean Absolute Error (MAE) is used to shown efficient of our proposed method.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.15
• /
• pp.6249-6254
• /
• 2015

Targeting breast cancer awareness along with comprehensive cancer care is appropriate in low and middle income countries like India, where there are no organized and affordable screening services. It is essential to identify the existing awareness about breast cancer in the community prior to launching an organized effort. This study assessed the existing awareness about breast cancer amongst women and their health seeking practices in an urban community in Mumbai, India. A postal survey was undertaken with low or no cost options for returning the completed questionnaires. The majority of the women were aware about cancer but awareness about symptoms and signs was poor. Women were willing to accept more information about cancer and those with higher awareness scores were more likely to seek medical help. They were also more likely to have undergone breast examination in the past and less likely to use alternative medicines. High income was associated with better awareness but this did not translate into better health seeking behaviour. Organized programmes giving detailed information about breast cancer and its symptoms are needed and women from all income categories need to be encouraged for positive change towards health seeking. Further detailed studies regarding barriers to health seeking in India are necessary.

• Journal of Korean Public Health Nursing
• /
• v.21 no.2
• /
• pp.241-251
• /
• 2007

Purpose: To identify the effects on tuberculosis mortality of a tuberculosis control program conducted at 108 community health centers in terms of structure and process. Methods: The dependent variable was tuberculosis mortality, and the independent variables were the structure(type of centers, staff, nurses, doctors, budget) and process(chest X-ray checking, immunization, case detection, health education, patients registering & managing) of the tuberculosis control programs at the community health centers. Data were analyzed using descriptive analysis and stepwise regression analysis. Result: Tuberculosis morality was positively correlated with type of centers(rural area)(p<0.01), but negatively correlated with type of centers(large cities) (p<0.01), (middle cities)(p<0.05), staff FTE(p<0.05), and number of nurses(p<0.05). Regression analysis indicated that type of centers(rural area)($\beta$=0.457) and case detection($\beta$=0.234) had a significant effect on tuberculosis mortality. Conclusion: Ultimately, this study will provide information to improve the effectiveness of tuberculosis control programs in community health centers.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.2
• /
• pp.559-564
• /
• 2015

Background: The participation of women in cervical cancer screening in Malaysia is low. Self-sampling might be able to overcome this problem. The aim of this study was to assess the reliability of self-sampling for cervical smear in our country. Materials and Methods: This cross-sectional study was conducted on 258 community dwelling women from urban and rural settings who participated in health campaigns. In order to reduce the sampling bias, half of the study population performed the self-sampling prior to the physician sampling while the other half performed the self-sampling after the physician sampling, randomly. Acquired samples were assessed for cytological changes as well as HPV DNA detection. Results: The mean age of the subjects was $40.4{\pm}11.3years$. The prevalence of abnormal cervical changes was 2.7%. High risk and low risk HPV genotypes were found in 4.0% and 2.7% of the subjects, respectively. A substantial agreement was observed between self-sampling and the physician obtained sampling in cytological diagnosis (k=0.62, 95%CI=0.50, 0.74), micro-organism detection (k=0.77, 95%CI=0.66, 0.88) and detection of hormonal status (k=0.75, 95%CI=0.65, 0.85) as well as detection of high risk (k=0.77, 95%CI=0.4, 0.98) and low risk (K=0.77, 95%CI=0.50, 0.92) HPV. Menopausal state was found to be related with 8.39 times more adequate cell specimens for cytology but 0.13 times less adequate cell specimens for virological assessment. Conclusions: This study revealed that self-sampling has a good agreement with physician sampling in detecting HPV genotypes. Self-sampling can serve as a tool in HPV screening while it may be useful in detecting cytological abnormalities in Malaysia.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.17 no.2
• /
• pp.75-82
• /
• 2017

In this paper, we consider social network analysis that focuses on community detection. Social networks embed community structure characteristics, i.e., a society can be partitioned into many social groups of individuals, with dense intra-group connections and much sparser inter-group connections. Exploring the community structure allows predicting as well as understanding individual's behaviors and interactions between people. In this paper, based on the interaction information extracted from a real-life Bluetooth contacts, we aim to reveal the social groups in a society of mobile carriers. Focusing on estimating the closeness of relationships between network entities through different similarity measurement methods, we introduce the clustering scheme to determine the underlying social structure. To evaluate our community detection method, we present the evaluation mechanism based on the basic properties of friendship.