Jiang, Liang-ChaoLiu, Run-RanJia, Chun-Xiao
7页查看更多>>摘要：Personalized recommender system is a powerful method to solve the problem of information overload, which has been widely applied in a variety of scenarios, such as e-commerce, video platforms and social networks, to help users find relevant items or friends of interest. Collaborative filtering is the most successful and widely used algorithm in the recommender systems as its powerful capability of generating recom-mendations by sharing collective experiences of users. In recent years, the use of mobile devices and the rapid development of internet infrastructures provide the possibility to analyze regional features of items based on user locations. Here we improve the performance of collaborative filtering by using user-location distribution to uncover the potential similarities between items. We find that the similarity of user-location distribution is one efficient measure for the item-item similarities in the framework of collaborative filtering to generate personalized recommendation for users. Furthermore, we have also mixed similarity measures of user-location distribution and the traditional method based on the number of common users linearly to optimize the performance of collaborative filtering. Based on the Movielens data set, we show that the performance of our methods could be improved in terms of the metrics of accuracy and diversity simultaneously. (C) 2021 Elsevier B.V. All rights reserved.
原文链接:
NSTL
Elsevier
Lotf, Jalil JabariAzgomi, Mohammad AbdollahiDishabi, Mohammad Reza Ebrahimi
13页查看更多>>摘要：Over the recent decade, much research has been conducted in the field of social networks. The structure of these networks has been irregular, complex, and dynamic, and certain challenges such as network topology, scalability, and high computational complexities are typically evident. Because of the changes in the structure of social networks over time and the widespread diffusion of ideas, seed sets also need to change over time. Since there have been limited studies on highly dynamical changes in real networks, this research intended to address the network dynamicity in the classical influence maximization problem, which discovers a small subset of nodes in a social network and maximizes the influence spread. To this end, we used soft computing methods (i.e., a dynamic generalized genetic algorithm) in social networks under independent cascade models to obtain a dynamic seed set. We modeled several graphs in a specified timestamp through which the edges and the nodes changed within different time intervals. Attempts were made to find influential individuals in each of these graphs and maximize individuals' influences in social networks, which could thereby lead to changes in the members of the seed set. The proposed method was evaluated using standard datasets. The results showed that due to the reduction of the search areas and competition, the proposed method has higher scalability and accuracy to identify influential nodes in these snapshot graphs as compared with other comparable algorithms. (C) 2021 Elsevier B.V. All rights reserved.
原文链接:- NSTL
- Elsevier
Assaf, AtaBilgin, Mehmet HuseyinDemir, Ender
14页查看更多>>摘要：In this paper, we use the transfer entropy to quantify information flows between three cryptocurrencies, namely Bitcoin, Ethereum and Ripple. We also employ the concept of Renyi transfer entropy that allows for capturing rare and frequent events separately as well as non-linear market dependencies, focusing on extreme (tail) observations of the return distributions. We find that Bitcoin and Ripple share a bidirectional information transmission, while there is only one directional information transmission from Ripple to Ethereum. There is no nonlinear information transmission according to the Renyi's measure, which implies the linear dependency among the three cryptocurrencies. This information transmission between cryptocurrencies occurs mostly in the pre-crash period while they become independent after the 2017 cryptocurrency crash. We finally use the concept of volatility surprise to examine linkages among the volatility of our series, and find a highly significant information transmission flow in one direction from Bitcoin to Ripple. Our results should be useful to investors in helping them in developing investment strategies by considering these three cryptocurrencies. (C) 2021 Elsevier B.V. All rights reserved.
原文链接:- NSTL
- Elsevier
Ruan, XiaoliXu, ChenFeng, JianwenWang, Jingyi...
15页查看更多>>摘要：This paper deals with the consensus problem of linear multi-agent systems (MASs) with matched uncertainties on directed topologies via a dynamic event-triggered control. To reduce the redundant communication, a distributed dynamic event-triggered (DET) protocol with a dynamic threshold is proposed, which makes the average time interval of controller update have a positive lower bound. The Zeno behavior can also be excluded. Furthermore, to avoid the assuming availability of full state information, an observer-based adaptive dynamic event-triggered (ADET) protocol is introduced. The proposed adaptive a-modification technique for the time-varying coupling gain can render smaller control gain and achieve better regulatory effects. Compared with the existing DET scheme, the dynamic threshold with a positive lower bound can reach consensus with larger inter-execution times and less communication energy among agents. By using the Riccati equation and inequality techniques, some simple and convenient sufficient conditions are derived to guarantee the stability of the closed-loop system. Finally, two numerical examples are given to verify the effectiveness of the obtained theoretical results. (C) 2021 Published by Elsevier B.V.
原文链接:- NSTL
- Elsevier
Lin, DanWu, JiajingChen, Jialan
15页查看更多>>摘要：Cryptocurrency exchanges, which act as a platform for cryptocurrency trading, play a vital role in the ever-growing cryptocurrency market. However, with the rapid development of this emerging market, some unethical phenomena including faking trading volume have also appeared in cryptocurrency exchanges. To this end, this paper proposes a data mining-based method based on off-chain data and on-chain transaction data to detect the exchanges that fake trading volume. In particular, we first collect off-chain data from the websites of five exchanges and the on-chain data provided by a blockchain browser, and then analyze them from two perspectives, including transaction number and transaction amount. The empirical results suggest that Huobi exchange fakes trading volume most obviously, while Binance trading is relatively the most honest. In addition, different exchanges adopt distinct counterfeiting strategies when creating wash trading. (C) 2021 Elsevier B.V. All rights reserved.
原文链接:- NSTL
- Elsevier
Dong, HanxuanDing, FanTan, HuachunZhang, Hailong...
16页查看更多>>摘要：Traffic prediction on a large-scale road network is of great importance to various applications. However, many factors such as sensor failure and communication errors inevitably resulted in a sparse distribution of effective detection points with missing data, which resulting adversely affects the accuracy of traffic prediction. This study considers the bidirectional connectivity of road networks to construct a two-way network graph topology. Based on the graph representation, the tensor combined temporal similarity revisited graph convolutional gate recurrent unit (T-TRGCGR), ingeniously combining traffic prediction and data completion through the Graph Laplace, is proposed to predict traffic states under partially input data missing circumstances and sparse detector distribution for a large-scale freeway network. Additionally, the proposed model can not only be applicable to traffic data prediction with missing values but also adaptively extract the spatio-temporal characteristics from various traffic periodicities while retaining the topological information of the large-scale network. Experiments on a large intercity network in Jiangsu, China shows that the proposed method outperforms state-of-art baselines on real-world traffic dataset, which can be well adapted to the prediction task of sparse coverage of road network detectors with missing data. Furthermore, through the comprehensive analysis and visualization of model parameters and results, it can be seen that the model adequately identifies the influential road network nodes and automatically learns to determine the importance of past traffic flow. (C) 2021 Elsevier B.V. All rights reserved.
原文链接:- NSTL
- Elsevier
Lu, JinquanChen, XinqiangYan, ZhongweiYan, Ying...
10页查看更多>>摘要：Ship trajectory from automatic identification system (AIS) provides crucial kinematic information for various maritime traffic participants (ship crew, maritime officials, shipping company, etc.), which greatly benefits the maritime traffic management in real-world. In that manner, ship trajectory smoothing and prediction attracts significant attentions in the maritime traffic community. To address the issue, an ensemble machine learning framework is proposed to remove outliers in the raw AIS data and predict ship trajectory variation tendency. Our method is verified on three typical ship trajectory segments, which is compared against other ship trajectory prediction models. The experimental results suggested that our proposed framework obtained higher prediction accuracy compared to the common trajectory prediction models in terms of typical error measurement indicators. The research findings can help maritime traffic participants obtain high-fidelity ship trajectory data, which supports making more reasonable traffic-controlling decisions. (C) 2021 Elsevier B.V. All rights reserved.
原文链接:- NSTL
- Elsevier
