228 research outputs found
Effects of climate variability on evaporation in Dongping Lake, China, during 2003–2010
No full textBased on two long-term, hourly (10:30–11:30 and 13:10–14:10) meteorological over-lake observations and data from Shenxian meteorological station, nearby Dongping Lake, the Penman-Monteith equation and reference evaporation ratio algorithm were used to calculate lake evaporation in Dongping Lake, China, from 2003 to 2010. The variation trend of evaporation of Dongping Lake was analyzed, and the influences that caused changes in lake evaporation were also discussed. The results show that (1) the total annual evaporation in Dongping Lake increased at 18.24 mm/a during 2003–2010.The major climatic factors accounting for this increase are the rising net radiation and the rising air temperature; (2) the total annual evaporation in a particular hour (13:10–14:10) in Dongping Lake increased at 4.55 mm/a during 2003–2010—the major climate factors that accounted for this increase are rising net radiation, followed by air temperature, wind velocity, and air humidity; (3) against the background of global warming, the climate of Dongping Lake tended to be dry during 2003–2010; the largest contribution to this comes from air temperature, followed by wind velocity and relative humidity; and (4) the monthly evaporation in Dongping Lake has seasonal variability.Water ManagementCivil Engineering and Geoscience
Does culture matter in effectiveness of social media marketing strategy? An investigation of brand fan pages
No full text
Purpose
Along with traditional marketing channels, social media outlets are integrated as a part of the marketing mix. Social media has changed the dynamics of interaction between companies and consumers that foster this relationship. Managing brand fan pages on social networking sites is a specific way the companies are using. Customers can become brand fans on these pages and indicate that they like the brand posts, share on their wall or simply comment. The purpose of this paper is to analyse the impact of cultural differences on the effectiveness of social media metrics and scientifically tested brand engagement in terms of commitment, loyalty and brand recommendations.
Design/methodology/approach
The authors analysed 1,922 brand posts from five different brands of a single product category in three different countries. Ordinary least square and hierarchical moderation regression was used to test the hypotheses.
Findings
Results show that all determinants are not equally suitable for enhancement of number of likes, comments and shares. More specifically, vivid and interactive brand post determinants enhance the number of likes. Furthermore, interactive brand posts enhance the number of comments while vivid brand posts enhance number of shares. Moreover, impact and intensity vary across different cultures.
Originality/value
Brand fan page moderators can obtain guidance from the research in formulating their social media marketing strategies in order to decide which post determinants to place on the fan page.
</jats:sec
Foucault’s governmentality and the issue of project collaboration
No full textThe concept of governmentality has proven useful to analyse how the reflexive management of people within and without the project is conducted. In this chapter we explore the organizational theory of governmentality and its importance in project settings. First, we identify the specificity of project governance and relate it to the definition and discussion of governmentality by Michel Foucault. Following this, the use of governmentality within projects through project culture is discussed. Subsequently, the use of governmentality outside projects through social media is discussed. Finally, the chapter concludes by highlighting new directions for research with governmentality as the focal point, discussing the types of research questions that a concern with projects and governmentality raises and how addressing these might further develop project management as a field of enquiry.Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.Integral Design & Managemen
Project narratives: Directions for research
No full textThis Research Handbook provides a cutting-edge review of complex project organizing (CPO), and suggests fruitful avenues for future research with a focus on grand challenges and a sustainable future.
Split into four sections, this Research Handbook addresses transitions within the field of CPO that could, and should, take place to achieve our shared aspirations for a better future. Featuring a team of contributors that is both interdisciplinary and geographically widespread, chapters provide a clarification of core concepts of complex project organizing, comprehensive coverage of leading theoretical perspectives for CPO, as well as a discussion of key empirical research themes. In particular, special attention is given to the implications of Industry 4.0 for complex project organizing.
The Research Handbook on Complex Project Organizing develops a guiding path to help academics – both established and early career – and research students in the fields of business leadership, operations management, and knowledge management navigate through these important topics, and envision how to respond to the grand challenges we all face
Warren Wilson College Political Science Professor to Lecture March 24
No full textDongping Han teaches political science at Warren Wilson College in Asheville, N.C. Han is the author of The Unknown Cultural Revolution: Life and Change in a Chinese Village
Emerg Infect Dis
No full textIn spring 2013, influenza A(H7N9) virus was isolated from an apparently healthy tree sparrow in Chongming Dongping National Forest Park, Shanghai City, China. The entire gene constellation of the virus is similar to that of isolates from humans, highlighting the need to monitor influenza A(H7N9) viruses in different species
Minimum-cost information dissemination in social networks
No full textSource type: Electronic(1) International Conference on Wireless Algorithms, Systems, and ApplicationsIn a social network, when the number of users discussing a topic exceeds a critical threshold, the topic will have a serious impact on the corresponding community. In this paper, we consider the problem of finding the minimum set of initial users of a topic to propagate a message so that, with a given guaranteed probability, the number of users discussing the topic would reach the critical threshold. This study is formally called the Minimum-Cost Information Dissemination (MCID) problem in our research. Different from the influence maximization problem, the MCID problem attempts to achieve influence maximization from the minimum cost perspective. To tackle the problem, we proposed a novel method based on h-hop independent set, HISS. Based on the independent set, HISS guarantees that the source nodes are sparsely distributed in the network. In addition, since HISS utilizes h-hop graph transformation, it can reduce the number of source nodes and avoid the scenarios in which the source nodes have common neighbors. The proposed method was evaluated with two real networks. The experimental results indicate that our proposed algorithm outperforms the state-of-the-art algorithms
Turbidite Fan Deposits in Gentle Slope Zones of Continental Faulted Basins: A Case Study from the Chezhen Depression, Bohai Bay Basin
No full textTurbidite fans, serving as good reservoirs for petroleum accumulation, are typically formed during deep faulting periods in continental basins, particularly in steep slope zones. However, gentle slope zones are also significant and unique for the formation of turbidite fans. These turbidite fans hold immense importance in exploring concealed lithological reservoirs. Taking the Chezhen Depression of Bohai Bay Basin as an example, we conducted a comprehensive study of the turbidite fan deposits in the gentle slope zone. Our results indicate that (1) small-scale distal-source turbidite fans are a common sedimentary type in the Chezhen Depression of the Bohai Bay Basin; (2) the study area is mainly characterized by seven lithofacies; (3) there are incomplete Bouma sequences in the study interval. This study is an important turbidite investigation into continental faulted basins, and it can also provide an important reference value for exploration and development in unconventional reservoirs of the same type
Applications of Novel Loss Functions and Interpretable Machine Learning for Clinical Risk Prediction
No full textClinical risk prediction focuses on the estimation of the probabilities of future adverse events in patients. Clinical risk prediction pipelines commonly involve different components such as data pre-processing, an appropriate model and loss (cost) function, imputation of missing values, assessment of discriminatory power and probability calibration, interpretability or explainability methods, and selection of a decision threshold. Although statistical models (such as logistic regression) have traditionally been used in clinical risk prediction pipelines, areas such as machine learning and computer vision have expanded the spectrum of loss functions, flexible models, and interpretable models that may be applicable to clinical risk prediction. An important question is whether these new loss functions and interpretable models can truly lead to better discriminatory power and probability calibration. This dissertation aims to improve clinical risk prediction pipelines through the application of novel loss functions for probability estimation and interpretable models that have emerged in the field of machine learning. The primary clinical application of this dissertation is the prediction of adverse events after thoracic organ transplantation (heart, lung, and heart-lung). These adverse events include rare malignancies, such as post-transplant lymphoproliferative disorder (PTLD), which increase the mortality rate of thoracic organ recipients after transplantation. The secondary application is the assessment of diabetes based on survey data. This dissertation provides two contributions to the domain of clinical risk prediction, prediction of transplant outcomes, and diabetes assessment:
[1] The first part of the dissertation proposes to train gradient-boosted decision trees (GBDT) using the focal loss function for clinical risk prediction tasks. Focal loss guides the model to focus on difficult or rare cases of patients during the training process. The properties of the focal loss function are explored in the context of probability estimation for clinical risk prediction. Focal-aware versions of the extreme gradient boosting machine (XGBoost) and the light gradient boosting machine (LightGBM) are developed to predict PTLD after lung transplantation. GBDTs trained using focal loss are also developed to evaluate the risk of diabetes using a publicly accessible diabetes dataset. Bayesian hyperparameter optimization is used to select the model hyperparameters that minimize the focal loss function. A closed-form expression that relates the true probability of the outcome to the focal loss minimizer is used to estimate calibrated probabilities. Decision thresholds are also explored to convert the probabilistic output of these models into a binary label to perform the PTLD classification. A comprehensive evaluation utilizing discrimination, probability calibration, and classification metrics is performed to determine utility in real-life clinical settings. We demonstrate that the focal-aware GBDT outperforms logistic regression and, in certain instances, surpasses the performance of the GBDT trained with the conventional cross-entropy loss function.
[2] The second part of the dissertation explores models with integer coefficients known as risk scores to predict the risk of PTLD in recipients of thoracic transplants (including heart, lung, and heart-lung). We adapt an evaluation technique known as rolling cross-validation to provide a reliable assessment of the entire modeling pipeline. Rolling cross-validation is also used to create a variable or feature importance plot that can help clinical professionals choose a risk score for real-life clinical settings. We show experiments where the risk scores achieve performance similar to less interpretable non-linear models when considering both discrimination and probability calibration metrics. The second part of the dissertation also seeks to improve the comprehension of the risk factors associated with PTLD through the application of these interpretable risk scores
THE DEVELOPMENT OF SYSTEMS SCIENCE AND ITS THREE CONTRIBUTIONS TO PHILOSOPHY OF SCIENCE
No full textFor a century since its inception in 1920s, systems science has constantly and consistently explored complexity, raising numerous new philosophical questions. Scholars have taken different approaches to understanding the development of systems science, which is of great significance to appreciating its philosophical value. In this regard, we concur with H. Simon that there have been three eruptions of interest in complexity and complex systems. We consider as the first eruption the research pioneered by A. Bogdanov and featuring Bertalanffy's general system theory, which marks the formation of systems science. The second eruption refers to the research of self-organization theory that developed from dissipative structure theory and synergetics. The third eruption refers to the research of complexity theory represented by complex adaptation system theory. We propose that there has formed a fourth eruption, as researchers now widely apply the concepts and methods of systems to economic, social and environmental systems, giving systems science research a strong application orientation. The advances in systems science have challenged some classical philosophical ideas and raised a new range of philosophical questions. Furthermore, the Elsevier Handbook of the Philosophy of Science incorporated Philosophy of Complex Systems into the framework of philosophy of science for the first time. We propose that the contributions of systems science to the philosophy of science manifests in the following three aspects: First, systems philosophy developed by Bertalanffy, including systems ontology, epistemology and value, has laid the foundation for systems philosophy to become a school of philosophy of science. Second, systems science addresses the relationship between whole and part, existence and evolution, and seeks to provide new justifications for ontological problems such as relational realism. Third, systems science also challenges the epistemological problems, carving out new explanatory approaches to complex systems emergence with condition-dependent laws and other principles of systems science
- …
