Search

The UK Biobank study contains several sources of diagnostic data, including hospital inpatient data and data on self-reported conditions for approximately 500,000 participants and primary-care data for approximately 177,000 participants (35%). Epidemiologic investigations require a primary disease definition, but whether to combine data sources to maximize statistical power or focus on only 1 source to ensure a consistent outcome is not clear. The consistency of disease definitions was investigated for venous thromboembolism (VTE) by evaluating overlap when defining cases from 3 sources: hospital inpatient data, primary-care reports, and self-reported questionnaires. VTE cases showed little overlap between data sources, with only 6% of reported events for persons with primary-care data being identified by all 3 sources (hospital, primary-care, and self-reports), while 71% appeared in only 1 source. Deep vein thrombosis–only events represented 68% of self-reported VTE cases and 36% of hospital-reported VTE cases, while pulmonary embolism–only events represented 20% of self-reported VTE cases and 50% of hospital-reported VTE cases. Additionally, different distributions of sociodemographic characteristics were observed; for example, patients in 46% of hospital-reported VTE cases were female, compared with 58% of self-reported VTE cases. These results illustrate how seemingly neutral decisions taken to improve data quality can affect the representativeness of a data set. [ABSTRACT FROM AUTHOR]

Currently, social media platforms become a valuable source of user-defined multi-features and multimedia data. Such data could be used for several event related applications, such as event detection, event relationship identification, and event representation in the form of nodes and edges. In this paper, we use a systematic critical review to assess recent works in event related problems to understand the state-of-the-art. Specifically, we mainly focus on the following recent major research issues: (i) event definition and the concept behind the 5W1H dimensions from social media perspective; (ii) event detection, including social media data representation, dimensions similarity measures, datasets, and evaluation measures used in the event detection task; (iii) event relationships identification covering the current approaches for three categories of event relationships: temporal, spatial, and semantic; and (iv) event-based knowledge graph construction in which details about event graph components and approaches used in the event graph are described. Furthermore, we provide applications, challenges, findings for each question, and promising research directions. [ABSTRACT FROM AUTHOR]

When considering the field of events, an important initial step is to consider the definition of an event. The goal of this article was twofold, first to take a novel approach in defining an event by considering the contextual aspects of an event from the traditional definitions, but adding the perspective of the attendee; and second to develop a typology focused on a unified definition of the field of events relational to the extant literature. The definition of an event is an occurrence that has a time element, two or more participants, is planned, and is a unique opportunity. A systematic review of extant literature was conducted in order to define the field of events. A total of 95 articles were analyzed. The field of events is characterized as having four major categories: professional, entertainment, social, and common cause. Within these four categories there are 20 types of events, all of which meet the four components of the definition of an event. [ABSTRACT FROM AUTHOR]

This paper presents a method to include damage at the initial stage of coastal flooding events definition and in return periods computation. The methodology is illustrated within a local study carried out in Biarritz Grande Plage, a meso-tidal, wave dominated beach located on the french basque coast in the south west of France. The work is based on two datasets covering the period 1949–2015: a first one, consisting of observation and synthetic data on wave characteristics and water level, and a second one, gathering storm dates and related damage intensities obtained through investigations in the press and in archives. A statistical analysis was first carried out to find the best combination of source variables explaining the reported damages for the identified storms. Maximal, mean and accumulated values were calculated over storm duration, considering source and aggregated variables based on the empirical run-up formula or the wave energy flux. Most rules combining a wave parameter and water level are found to provide satisfactory damage prediction as soon as maxima variables are considered. Rules based on mean variables are less accurate and those based on accumulated variable values are not relevant. The ability of the rules to be used as generic event definition rules is then tested by performing a retrospective analysis on the whole dataset, checking their efficiency in detecting historical storms (i.e., with damages) without finding too many false positives. Most of the rules formerly studied, except the ones using wave period only as wave parameter, were able to correctly perform this task. Coastal flood event return periods (RP) were then calculated by applying three of the best rules identified previously. The rule using non simultaneous maxima of wave energy flux and water level gives encouraging results for the RP values. Nevertheless, the discrepancy still observed among the different rules calls for further work in this direction. • Statistical link between historical damages due to waves and water level data studied. • Rules combining wave and water level maxima over storm duration are good damage predictors. • The best rules can be used blind as damaging events tracking. • Link between damage and return period: encouraging results obtained with the best rule. • Still large discrepancy among the rule for individual event return period quantification. [ABSTRACT FROM AUTHOR]

A normative description of a motion details the necessary and sufficient criteria to identify that motion. It equips researchers with a shared lexicon for describing their research and, in this way, adoption of a normative description facilitates communication within the research community. Although there is an abundance of descriptions of sit-to-stand movement, there is not a commonly accepted normative description of sit-to-stand; study-specific descriptions are commonplace. This work evaluates the breadth of existing sit-to-stand descriptions using new experimental data from 15 healthy young adults standing from a 46 cm chair. Our goal is to develop a normative description of the sit-to-stand motion that is in harmony with the literature. After aligning experimental data to seat-off (the one sit-to-stand event with a clear definition), events defining the start of sit-to-stand, seat-off, and the end of sit-to-stand are identified using a density-based clustering method. Then, the intermediary events of start of seat unloading, end of momentum transfer, and beginning of stabilization are determined while maintaining consistent sequencing and biomechanical meaning. These six events of sit-to-stand are determined from trunk, hip, knee, and ankle angle data and vertical ground reaction forces. The events are in greatest accordance with the descriptions of sit-to-stand introduced by Schenkman et al. (1990) and Kralj et al. (1990) , and the event timings are in alignment with the findings of other researchers. The proposed description of healthy sit-to-stand promotes consistency in the description of this motion and adoption of this description will promote effective communication in sit-to-stand research. [ABSTRACT FROM AUTHOR]

Efficient network monitoring is an important basis work for network management. Generally, many management applications require accurate and timely statistics about network states at different aggregation levels at low cost, such as malicious traffic detection, traffic engineering, etc. Moreover, the network environment to be monitored is constantly changing and expanding, including not only the data center for cloud computing but also the Internet of Things (IoT) for smart urban sensing, which requires the intensive study of more fine-grained network monitoring. As is well known, the development of efficient network monitoring approaches greatly relies on a flexible monitoring framework. Software defined network (SDN) can provide dramatic advantages for network management by separating the control plane and data plane. Therefore, it is a good choice to design a flexible monitoring framework based on the advantages of SDN. However, most research works only take advantage of the centralized control feature in SDN, which leads to limited improvement in the flexibility of the monitoring framework. This paper proposes a flexible monitoring framework named FlexMonitor, which can realize greater flexibility based on not only the centralized control feature, but also the high programmability in the controller and the limited programmability in the openflow switches in SDN. There are two key parts in FlexMonitor, namely the monitoring strategy deployment part and the monitoring data collection part, which can enrich the deployment methods of monitoring strategies and increase the kinds of monitoring data sources, respectively. Based on the NetMagic platform, this monitoring framework was implemented and evaluated through realizing a distributed denial of service (DDoS) detection approach. The experimental results show that the proposed DDoS detection approach has a better detection performance compared with other related approaches as well as indirectly show that FlexMonitor can flexibly support a variety of efficient monitoring approaches. [ABSTRACT FROM AUTHOR]