Search

Tourism efficiency is crucial for measuring sustainable tourism development. Examining the relationship between aviation and tourism efficiency networks is key to promoting their synergistic development in China's urban areas. This study employs various methods, such as complex network analysis method, entropy-weighted TOPSIS, tourism efficiency gravity model, and quadratic assignment procedure, to analyze the networks' spatial structure evolution characteristics and interaction effects. Results show that (1) China's major cities' aviation network has improved its organizational efficiency and formed a "double rhombus-single axis" spatial evolution pattern of the axis-spoke network. The number of intermediary networks and hub cities in the central and western regions has increased. (2) The tourism efficiency network adopts a "honeycomb" structure pattern with the simultaneous layout of "point-to-point" and "star" networks. The network's tourism efficiency follows "Pareto's Law," and tourism cities above the second level form a club group development. The tourism efficiency development potential area is shifting to the southwest. (3) The aviation and tourism efficiency networks exhibit a clear trend of synergistic evolution with a "path locking" phenomenon between them. Differences in tourism resource endowment, labor advantage, and capital advantage positively impact the aviation network's structure. Conversely, differences in revenue capacity and market scale negatively impact the structure. The aviation scale advantage, openness, intimacy, and influence exhibit decreasing positive effects on the tourism efficiency network's structure.

Sequential recommendation can make predictions by fitting users’ changing interests based on the users’ continuous historical behavior sequences. Currently, many existing sequential recommendation methods put more emphasis upon users’ recent preference (i.e., short-term interests), but simplify or even ignore the influence of users’ long-term interests, resulting in important interest features of users not being effectively mined. Moreover, users’ real intentions may not be fully captured by only focusing on their behavior histories, because users’ interests are diverse and dynamic. To solve the above problems, we propose a novel sequential recommendation model for long-term interest memory and nearest neighbor influence. Firstly, item embeddings based on item similarity and dependency are constructed to alleviate the problem of data sparsity in users’ recent interest history. Secondly, in order to effectively capture long-term interests, the long sequence is divided into multiple nonoverlapping subsequences. For these subsequences, the graph attention network with node importance factor is designed to fully extract the main interests of subsequences, and LSTM is introduced to learn the dynamic changes of interest among subsequences. Long-term interests of users are modeled through complex structure within subsequences and sequential dependencies among subsequences. Finally, the user’s neighbor representation is introduced, and a gating module is designed to integrate the user’s neighbor information and self-interests. The influence of users’ short-term and long-term interests on prediction is dynamically controlled by considering nearby features in the gating network. The experimental results on two public datasets show that the proposed sequential recommendation model can outperform the baseline methods in hit rate (HR@K) and normalized discounted cumulative gain (NDCG@K).

To protect recommender systems against shilling attacks, a variety of detection methods have been proposed over the past decade. However, these methods focus mainly on individual features and rarely consider the lockstep behaviours among attack users, which suffer from low precision in detecting group shilling attacks. In this work, we propose a three-stage detection method based on strong lockstep behaviours among group members and group behaviour features for detecting group shilling attacks. First, we construct a weighted user relationship graph by combining direct and indirect collusive degrees between users. Second, we find all dense subgraphs in the user relationship graph to generate a set of suspicious groups by introducing a topological potential method. Finally, we use a clustering method to detect shilling groups by extracting group behaviour features. Extensive experiments on the Netflix and sampled Amazon review datasets show that the proposed approach is effective for detecting group shilling attacks in recommender systems, and the F1-measure on two datasets can reach over 99 percent and 76 percent, respectively.

Collaborative recommender systems are vulnerable to shilling attacks. To address this issue, many methods including supervised and unsupervised have been proposed. However, supervised detection methods require training classifiers and they only apply to detect known types of attacks. The existing unsupervised detection methods need to know the prior knowledge of attacks, otherwise they suffer from low detection precision. In this paper, we present BS-SC, an unsupervised approach for detecting shilling profiles, which does not need to know the attack size or to label the candidate spammers. BS-SC starts from an in-depth analysis of user behaviors and uses two key mechanisms (i.e., behavior features extraction and behavior similarity matrix clustering) to distinguish shilling profiles from genuine ones. The behavior features reflect the behavior difference between genuine and shilling profiles, and the behavior similarity matrix clustering is to cluster shilling profiles based on their highly similar behaviors. Experimental results on the MovieLens and the sampled Amazon review datasets indicate that BS-SC outperforms the baseline unsupervised approaches, even when the prior knowledge is given for them.

Sum-frequency generation (SFG) vibrational spectroscopy is a second-order nonlinear optical spectroscopy technique. Owing to its interfacial selectivity, SFG vibrational spectroscopy can provide interfacial molecular information, such as molecular orientations and order, which can be obtained directly, or molecular density, which can be acquired indirectly. Interfacial molecular behaviors are considered the basic factors for determining the tribological properties of surfaces. Therefore, owing to its ability to detect the molecular behavior in buried interfaces in situ and in real time, SFG vibrational spectroscopy has become one of the most appealing technologies for characterizing mechanisms at friction interfaces. This paper briefly introduces the development of SFG vibrational spectroscopy and the essential theoretical background, focusing on its application in friction and lubrication interfaces, including film-based, complex oil-based, and water-based lubricating systems. Real-time detection using SFG promotes the nondestructive investigation of molecular structures of friction interfaces in situ with submonolayer interface sensitivity, enabling the investigation of friction mechanisms. This review provides guidance on using SFG to conduct friction analysis, thereby widening the applicability of SFG vibrational spectroscopy.

In this paper, we introduce a new setting for graph embedding, which considers embedding communities instead of individual nodes. We find that community embedding is not only useful for community-level applications such as graph visualization but also provide an exciting opportunity to improve community detection and node classification. Specifically, we consider the interaction between community embedding and detection as a closed loop, through node embedding. On the one hand, node embedding can improve community detection since the detected communities are used to fit a community embedding. On the other hand, community embedding can be used to optimize node embedding by introducing a community-aware high-order proximity. However, in practice, the number of communities can be unknown beforehand; thus we extend our previous Community Embedding (ComE) model. We propose ComE+, a new model which handles both: the unknown truth community assignments and the unknown number of communities present in the dataset. We further develop an efficient inference algorithm for ComE+ for keeping complexity low. Our extensive evaluation shows that ComE+ improves the state-of-the-art baselines in various application tasks, e.g., community detection and node classification.

With the development of social networks, the research of integrated social information recommendation models has received extensive attention. However, most existing social recommendation models are based on the matrix factorization technique which ignore the impact of the relationships between items on users’ interests, resulting in a decline of recommendation accuracy. To solve this problem, this paper proposes a double regularization matrix factorization recommendation algorithm. The algorithm first uses attribute information and manifold learning to calculate similarity. Then, the matrix factorization model is constrained through the regularization of item association relations and user social relations. Experimental results on real datasets show that the proposed method can effectively alleviate problems such as cold start and data sparsity in the recommender system and improve the recommendation accuracy compared with those of existing methods.

Most existing community-related studies focus on detection, which aim to find the community membership for each user from user friendship links. However, membership alone, without a complete profile of what a community is and how it interacts with other communities, has limited applications. This motivates us to consider systematically profiling the communities and thereby developing useful community-level applications. In this paper, we for the first time formalize the concept of community profiling. With rich user information on the network, such as user published content and user diffusion links, we characterize a community in terms of both its internal content profile and external diffusion profile. The difficulty of community profiling is often underestimated. We novelly identify three unique challenges and propose a joint Community Profiling and Detection (CPD) model to address them accordingly. We also contribute a scalable inference algorithm, which scales linearly with the data size and it is easily parallelizable. We evaluate CPD on large-scale real-world data sets, and show that it is significantly better than the state-of-the-art baselines in various tasks., Technical report of a PVLDB 2017 paper

Network embedding has been intensively studied in the literature and widely used in various applications, such as link prediction and node classification. While previous work focus on the design of new algorithms or are tailored for various problem settings, the discussion of initialization strategies in the learning process is often missed. In this work, we address this important issue of initialization for network embedding that could dramatically improve the performance of the algorithms on both effectiveness and efficiency. Specifically, we first exploit the graph partition technique that divides the graph into several disjoint subsets, and then construct an abstract graph based on the partitions. We obtain the initialization of the embedding for each node in the graph by computing the network embedding on the abstract graph, which is much smaller than the input graph, and then propagating the embedding among the nodes in the input graph. With extensive experiments on various datasets, we demonstrate that our initialization technique significantly improves the performance of the state-of-the-art algorithms on the evaluations of link prediction and node classification by up to 7.76% and 8.74% respectively. Besides, we show that the technique of initialization reduces the running time of the state-of-the-arts by at least 20%., Full Research Paper accepted in the 13th ACM International Conference on Web Search and Data Mining (WSDM 2020)

Tweet streams provide a variety of real-life and real-time information on social events that dynamically change over time. Although social event detection has been actively studied, how to efficiently monitor evolving events from continuous tweet streams remains open and challenging. One common approach for event detection from text streams is to use single-pass incremental clustering. However, this approach does not track the evolution of events, nor does it address the issue of efficient monitoring in the presence of a large number of events. In this paper, we capture the dynamics of events using four event operations (create, absorb, split, and merge), which can be effectively used to monitor evolving events. Moreover, we propose a novel event indexing structure, called Multi-layer Inverted List (MIL), to manage dynamic event databases for the acceleration of large-scale event search and update. We thoroughly study the problem of nearest neighbour search using MIL based on upper bound pruning, along with incremental index maintenance. Extensive experiments have been conducted on a large-scale real-life tweet dataset. The results demonstrate the promising performance of our event indexing and monitoring methods on both efficiency and effectiveness.

Graph is an important data representation which appears in a wide diversity of real-world scenarios. Effective graph analytics provides users a deeper understanding of what is behind the data, and thus can benefit a lot of useful applications such as node classification, node recommendation, link prediction, etc. However, most graph analytics methods suffer the high computation and space cost. Graph embedding is an effective yet efficient way to solve the graph analytics problem. It converts the graph data into a low dimensional space in which the graph structural information and graph properties are maximally preserved. In this survey, we conduct a comprehensive review of the literature in graph embedding. We first introduce the formal definition of graph embedding as well as the related concepts. After that, we propose two taxonomies of graph embedding which correspond to what challenges exist in different graph embedding problem settings and how the existing work address these challenges in their solutions. Finally, we summarize the applications that graph embedding enables and suggest four promising future research directions in terms of computation efficiency, problem settings, techniques and application scenarios., A 20-page comprehensive survey of graph/network embedding for over 150+ papers till year 2018. It provides systematic categorization of problems, techniques and applications. Accepted by IEEE Transactions on Knowledge and Data Engineering (TKDE). Comments and suggestions are welcomed for continuously improving this survey

The exponential growth of online videos, along with increasing user involvement in video-related activities, has been observed as a constant phenomenon during the last decade. User's time spent on video capturing, editing, uploading, searching, and viewing has boosted to an unprecedented level. The massive publishing and sharing of videos has given rise to the existence of an already large amount of near-duplicate content. This imposes urgent demands on near-duplicate video retrieval as a key role in novel tasks such as video search, video copyright protection, video recommendation, and many more. Driven by its significance, near-duplicate video retrieval has recently attracted a lot of attention. As discovered in recent works, latest improvements and progress in near-duplicate video retrieval, as well as related topics including low-level feature extraction, signature generation, and high-dimensional indexing, are employed to assist the process. As we survey the works in near-duplicate video retrieval, we comparatively investigate existing variants of the definition of near-duplicate video, describe a generic framework, summarize state-of-the-art practices, and explore the emerging trends in this research topic.

With the prevalence of intelligent devices, social networks have been playing an increasingly important role in our daily life. Various social networks (e.g., Twitter, Facebook) provide convenient platforms for users to explore the world. In this thesis, we study the problem of multi-perspective analysis of social events detected from social networks. In particular, we aim to make sense of the social events from the following three perspectives: 1) what are these social events about; 2) how do these events evolve along timeline; 3) who are involved in the discussions on these events. We mainly work on two categories of social data: the user-generated contents such as tweets and Facebook posts, and the users' interactions such as the follow and reply behaviours among users. On one hand, the posts reveal valuable information that describes the evolutions of miscellaneous social events, which is crucial for people to understand the world. On the other hand, users' interactions demonstrate users' relationships among each other and thus provide opportunities for analysing the underlying communities behind the social events. However, it is not practical to manually detect social events, monitor event evolutions or profile the underlying communities from the massive amount of social data generated everyday. Hence, how to efficiently and effectively extract, manage and analyse the useful information from the social data for multi-perspective social events understanding is of great importance. The social data is dynamic source of information which enables people to stay informed of what is happening now and who are the active and influential users discussing these social events. For one thing, social data is generated by people worldwide at all time, which may make fast identification of events even before the mainstream media. Moreover, the continuous stream of social data reflects the event evolutions and characterizes the events with changing opinions at different stages. This provides an opportunity to people for timely responses to urgent events. For another, users are often not isolated in social networks. The interactions between users can be utilized to discover the communities who discuss each social event. Underlying community profiling provides answers to the questions like who are interested in these events, and which group of people are the most influential users in spreading certain event topics. These answers deepen our understanding of the social events by considering not only the events themselves but also the users behind these events. The first research task in this thesis is to monitor and index the evolving events from social textual contents. The social data cover a wide variety of events which typically evolve over time. Although event detection has been actively studied, most existing approaches do not track the evolution of events, nor do they address the issue of efficient monitoring in the presence of a large number of events. In this task, we detect events based on the user-generated textual contents and design four event operations to capture the dynamics of events. Moreover, we propose a novel event indexing structure, called Multi-layer Inverted List, to manage dynamic event databases for the acceleration of large-scale event search and update. The second research task is to explore multiple features for social events tracking and visualization. In addition to textual contents utilized in the first task, social data contains various features, such as images and timestamps. The benefits of incorporating different features into event detection are twofold. First, these features provide supplemental information that facilitates the event detection model. Second, different features describe the detected events from different aspects, which enables users to have a better understanding with more vivid visualizations. To improve the event detection performance, we propose a novel generative probabilistic model which jointly models five different features. The event evolution tracking is achieved by applying the maximum-weighted bipartite graph matching on the events discovered in consecutive periods. Events are then visualized by the representative images selected based on our three defined criteria. The third research task is to detect and profile the underlying social communities in social events. The social data not only contains user-generated contents which describe the events evolutions, but also comprises various information on the users who discuss these events, such as user attributes, user behaviours, and so on. Comprehensively utilizing this user information can help to group similar users into communities, and enrich the social event analysis from the community perspective. Motivated by the rich semantics about user behaviours hidden in social data, we extend the community definition as a group of users who are not only densely connected, but also having similar behaviours. Moreover, in addition to detecting the communities, we further profile each of the detected communities for social events analysis. A novel community profiling model is designed to detect and characterize a community by both content profile (what a community is about) and diffusion profile (how it interacts with others).