Multi-attribute Event Research Articles

Bivariate copula for flood frequency analysis in Johor river basin

Flooding is a multi-attribute event that is described by many factors such as peak flow and flood volume. It is extremely vital to consider both the flood volume and the flood peak while studying the flood frequency analysis as the univariate analysis cannot accurately portray the flood issue and suffers from an underestimation and an overestimation problem. Traditional univariate and multivariate modeling techniques have several mathematical shortcomings including the inability to distinguish between the marginal and joint behavior of the variables under study. Therefore, the copula function was introduced to tackle the above restriction. Six copula models will be applied in this study to find the best bivariate distribution between the flood variables in Johor River Basin, Malaysia, including Gaussian, Student-t, Clayton, Gumbel, Frank, and Joe. Before that, several marginal distributions were fitted to select the most appropriate distribution for flood variables. It was found that the Pearson Type-III fits both the flood peak flow and the flood volume best. The evaluation of the best univariate marginal distribution and the copula model will be based on Akaike Information Criterion (AIC). Our findings suggested that Frank Copula is more suited to represent the relationship between peak flow and flood volume as it portrays the lowest AIC values of -69.41 and highest log-likelihood values of 35.7, where both values outperform the other proposed copula models. However, future research which considers all three flood variables which are peak flow, volume, and duration should be conducted to attain a more reliable result.

A Dempster‐Shafer theory and uninorm‐based framework of reasoning and multiattribute decision‐making for surveillance system

Closed-circuit television and sensor-based intelligent surveillance systems have attracted considerable attentions in the field of public security affairs. To provide real-time reaction in the case of a huge volume of the surveillance data, researchers have proposed event-reasoning frameworks for modeling and inferring events of interest. However, they do not support decision-making, which is very important for surveillance operators. To this end, this paper incorporate a function of decision-making in an event-reasoning framework, so that our model not only can perform event-reasoning but also can predict, rank, and alarm threats according to uncertain information from multiple heterogeneous sources. In particular, we propose a multiattribute decision-making model, in which an object being watched is modeled as a multiattribute event, where each attribute corresponds to a specific source, and the information from each source can be used to elicit a local threat degree of different malicious situations with respect to the corresponding attribute. Moreover, to assess an overall threat degree of an object being observed, we also propose a method to fuse the conflict threat degrees regarding all the relevant attributes. Finally, we demonstrate the effectiveness of our framework by an airport security surveillance scenario.

Learning from major accidents: Graphical representation and analysis of multi-attribute events to enhance risk communication

Major accidents are complex, multi-attribute events, originated from the interactions between intricate systems, cutting-edge technologies and human factors. Usually, these interactions trigger very particular accident sequences, which are hard to predict but capable of producing exacerbated societal reactions and impair communication channels among stakeholders. Thus, the purpose of this work is to convert high-dimensional accident data into a convenient graphical alternative, in order to overcome barriers to communicate risk and enable stakeholders to fully understand and learn from major accidents. This paper first discusses contemporary views and biases related to human errors in major accidents. The second part applies an artificial neural network approach to a major accident dataset, to disclose common patterns and significant features. The complex data will be then translated into 2-D maps, generating graphical interfaces which will produce further insight into the conditions leading to accidents and support a novel and comprehensive “learning from accidents” experience.

DEXIN: A fast content-based multi-attribute event matching algorithm using dynamic exclusive and inclusive methods

Event matching acts as a key role in publish/subscribe services, which takes charge of finding all subscriptions that match an event from a subscription set. In large content-based multi-attribute publish/subscribe systems, the performance of event matching is severely challenged due to the increased system scale and search complexity. Most existing event matching algorithms cannot effectively sustain in such a scenario. In this paper, we propose DEXIN (Dynamic EXclusive and INclusive), a fast content-based multi-attribute event matching algorithm using dynamic exclusive and inclusive methods. DEXIN supports highly efficient event matching for subscriptions with limited range-based, unlimited range-based and point-based constraints. DEXIN prepares a pipelined event matching process optimized for each incoming event. The exclusive and inclusive methods, which have different matching costs, are dynamically chosen for event matching over a single attribute. Thus, the search space shrinks very fast since non-matched subscriptions can be filtered out by DEXIN at very early stages in the pipeline. The experiment results demonstrate that DEXIN is superior to several state-of-the-art reference algorithms. Especially, the leading advantage of DEXIN is more significant on matching time and stability for large number of subscriptions and attributes in multiple scenarios.

Toward high efficiency for content-based multi-attribute event matching via hybrid methods

Event matching is a core in decoupled end-to-end communications, which are extensively applied to various areas. Event matching seeks the subscriptions that match a given event from a subscription set, however, this work becomes increasingly complicated in content-based multi-attribute scenarios, where events and subscriptions are formed in content, and described by multiple attributes. In addition, large-scale systems are easier to suffer from severe degradation in event matching performance. To this end, this paper presents a high-efficiency content-based multi-attribute event matching algorithm, called HEM (hybrid event matching), which is hybridized by 2 different methods. In HEM, the matching on each single attribute (called single-attribute matching) is processed by a triangle-based matching method or a direct matching method dynamically. All single-attribute matchings are sorted via a fast near-optimal algorithm, and each of them is carried out sequentially. In this manner, the searching space of event matching shrinks gradually, so that the searching performance is boosted along with the process of event matching. Experiments are conducted to evaluate HEM comprehensively, where it is observed that HEM outperforms 3 state-of-the-art counterparts (TAMA, H-TREE and REIN) in main criteria, such as event matching time, insertion time and deletion time. Moreover, the gap of performance between HEM and the counterparts enlarges with the increase of system scale.

Learning probabilistic real-time automata from multi-attribute event logs

The growing number of time-labeled datasets in science and industry increases the need for algorithms that automatically induce process models. Existing methods are capable of identifying process models that typically only work on single attribute events. We propose a new model type to address the problem of mining multi-attribute events, meaning that each event is described by a vector of attributes. The model is based on timed automata, includes expressive descriptions of states and can be used for making predictions. A probabilistic real time automaton is created, where each state is annotated by a profile of events. To identify the states of the automaton, similar events are combined by a clustering approach. The method was implemented and tested on a synthetic, a medical and a biological dataset. Its prediction accuracy was evaluated on a medical dataset and compared to a combined logistic regression, which is considered a standard in this application domain. Moreover, the method was experimentally compared to Multi-Output HMMs and Petri nets learned by standard process mining algorithms. The experimental comparison suggests that the automaton-based approach performs favorably in several dimensions. Most importantly, we show that meaningful medical and biological process knowledge can be extracted from such automata.

An architecture for monitoring, visualization, and control of gigabit networks

We outline a design of a highly scalable network monitoring, visualization, and control (NMVC) system with advanced algorithmic and human-in-the-loop capability. This capability allows network administrators to calibrate and fine-tune network and application parameters in real time according to observed traffic patterns. The goal of the NMVC system is to ensure adequate quality of service to network users, while maintaining high network resource utilization. The main components of our system are: a network probe and an endsystem probe which can probe gigabit/s links, software network management agents that provide extensible multi-attribute event filtering for highly scalable date/event collection, efficient online event ordering algorithms that can help synthesize and display a consistent view of network health, status, and performance and a View Choreographer that allows management applications and administrators to specify the mapping of network events to higher-level events and to visualization objects and updates.