Abstract This paper proposes an attention-based LSTM (AT-LSTM) model for financial time series prediction. We divide the prediction process into two stages. For first stage, we apply attention to assign different weights input features of at each step. In second feature is utilized effectively select relevant sequences as neural network in next frame. Our proposed framework not only solves long-term dependence problem effectively, but also improves interpretability methods based on network....

Abstract In this work, we investigate the issue of inadequate solution quality and insufficient convergence in multi-sensor scheduling algorithms for tracking aerial moving targets, enhance traditional Non-dominated Sorting Genetic Algorithm II (NSGA-II) through adaptive parameters crossover mutation. Taking error sensor resource consumption into account, establish two task metrics a multi-objective optimization problem. The refined algorithm is employed to solve Simulation results indicate...

In the research of social networks, their structural characteristics, user behaviors, and relationships are elements that important to understand predict manage networks. this article, we took as object networks mobile communication network, which is closely connected with people's real lives. We studied characteristics using complex network analysis methods derived laws exist in structure analyzed users' behaviors or patterns from perspective age, gender, scope, age differences, time....

Unsupervised multi-modal hashing has recently attracted broad attention in research area of large-scale multimedia retrieval for its low storage cost, high speed, and independence on semantic labels. However, the model learning process existing methods still suffer from problem efficiency: 1) Many measure contributions different modalities using fixed modality weights. In order to avoid over-fitting, they need an inefficient hyper-parameter adjustment process. 2) Most adopt optimization...

Paths and cycles are the two pivotal elements in a network. Here, we demonstrate that paths, particularly shortest ones, incomplete information However, based on such many network centrality measures designed. While extensive explorations paths have been made, modest studies focus measuring centrality. We study relationship between cycle path from real-world networks. The results illustrate incompleteness of Noticing is much more robust than path, propose novel cycle-based to address paths:...

Abstract The description of user behavior in social networks is an important issue for studying networks. Given that Petri nets can describe the resource flow problem, this study utilizes features to portray states during message propagation a micro-blog network and presents information formalized representation method network. On basis, analyzed proposed detail. We provide examples applying (e.g., addiction users, influence, public opinion analysis). In addition, we introduce algorithms...

Social network user identification is of great value and significance for predicting users' behavior, analyzing regulating social networks, studying the interaction between users. This paper studies identities users, divides them into organizational individual users in terms their identities, explicitly defines identifies both types also distinguishes a as an organization or according to contents text, multimedia, time series published network. During identification, content (topic)...

In order to carry on a backlog and deliver the knowledge of management experience, means, organization circulate way in each stage whole life span process management, equipment system is constructed integrate support platform, repository, application exchanges platform expert answer on-line cooperation tool making use military local area network. Therefore, has function information releases, search, forum answer, study, real time cooperation, feedback etc., for users work provides one-stop...

This article has been retracted by the publisher.

In order to solve the problem that measurement noise covariance of navigation components in integrated system unmanned aerial vehicle (UAV) close formation flight is time-varying or difficult detect accurately, which leads decline filtering accuracy, zero-reset federated filter adopted. Different calculation periods and fusion are designed fuse information with unequal intervals. Based on idea variational Bayesian inference subfilter, real posterior distribution approximated by a simple...

Information retrieval in real-time search presents unique challenges distinct from those encountered classical web search. These are particularly pronounced due to the rapid change of user intent, which is influenced by occurrence and evolution breaking news events, such as earthquakes, elections, wars. Previous dense methods, primarily focused on static semantic representation, lack capacity capture immediate leading inferior performance retrieving most recent event-related documents...

Abstract Under the background of “cultural and tourism integration”, scenic cultural creative experience products are an important means to enhance communication power spots generate income. In this paper, CORE model is used as design concept, process for constructed. The plain Bayesian classification algorithm extract review feedback from product, vectorization extracted reviews achieved using FastText model. vectorized input into logistic regression obtain sentiment results. A fused...

It is an important core issue in social networks to divide community or group.And, the network node mainstream of division algorithm as processing object network.This paper introduces entropy theory into partition on basic studying concept and algorithm.And we proposed Entropy-based Link Partition (ELP algorithm), which links object.Also, similarity between two objects properly defined improved, thus more close real situation network.Experimentation real-world networks, obtained results...