HE-SNE: Heterogeneous Event Sequence-based Streaming Network Embedding for Dynamic Behaviors

User interest and behavior modeling is a critical step in online digital advertising. On the one hand, user interests directly impact their response and actions to the displayed advertisement (Ad). On the other hand, user interests can further help determine the probability of an Ad viewer becoming a buying customer. To date, existing methods for Ad click prediction, or click-through rate prediction, mainly consider representing users as a static feature set and train machine learning classifiers to predict clicks. Such approaches do not consider temporal variance and changes in user behaviors, and solely rely on given features for learning. In this paper, we propose two deep learning-based frameworks, {hbox {LSTM}}_{mathrm{cp}} and {hbox {LSTM}}_{mathrm{ip}}, for user click prediction and user interest modeling. Our goal is to accurately predict (1) the probability of a user clicking on an Ad and (2) the probability of a user clicking a specific type of Ad campaign. To achieve the goal, we collect page information displayed to the users as a temporal sequence and use long short-term memory (LSTM) network to learn features that represents user interests as latent features. Experiments and comparisons on real-world data show that, compared to existing static set-based approaches, considering sequences and temporal variance of user requests results in improvements in user Ad response prediction and campaign specific user Ad click prediction.

In the era of digitization, online digital advertising is one of the best techniques for modern marketing. This makes advertisers rely heavily on accurate user interest and behavior modelling to deliver precise advertisement impressions and increase click-through rates. The classic approach to model user interests has often required the use of predefined feature sets which are typically stagnant and not representative of temporal changes and trends in user behavior. While recent advances in deep learning offer potential solutions to these obstacles, many existing approaches fail to address the sequential nature of user interactions. In this paper, we propose an optimized Bi-Directional Long Short-Term Memory (Bi-LSTM) based user interest modeling approach together with an Updated version of Parrot Optimizer (UPO) to enhance performance. It treats the user activity as a temporal sequence which well fits the changing nature of user interest and preferences over time. The proposed approach is evaluated on two important tasks: predicting the probability that a user will click on an ad and predicting the probability that a user will click on a particular type of ad campaign. Simulation results demonstrate that the proposed method provides superior results than the static set-based approaches and achieves significant improvements on both user ad responses predictions and user ad clicks at the campaign level. The research also enhances the efficiency of user interest modeling with significant implications for online advertising, recommendation systems, and personalized marketing.

Finding prevalent mobile user patterns in large amount of data has been one of the major problems in the area of mobile data mining. Particularly, the algorithms of discovering frequent user's behavior patterns in the mobile agent system have been studied extensively in recent years. The key feature in most of these algorithms is that they use a location log dataset with userpsilas requested services. In this case, some problems occur because they do not consider that mobile user's behavior patterns are dynamically variable as time passes. In this paper, we propose a novel data mining method based on temporal mobile access patterns that can efficiently discover mobile user's behavior patterns. An advantage of our method compactly stores user's behavior patterns according to location log dataset and requested services in memory. Thus, even the information datasets require large shared memory when they store, our approach still provides faster access and consume less memory than existing techniques.

Users for online shopping are increasing day by day because of easy to get and time-saving property of online shopping. Having a proper understanding of users interest for certain type of product or different products for online shopping becomes important to create personalized service for a target market. An important property of successful e-commerce website is the ability to provide useful content at the right time to users. And because of all this, personalization techniques are introduced to create adaptive shopping application in which user interfaces change according to users interest. User’s behavior information is stored in web log files, and to get the information data mining techniques are used in which they use statistical characters to model users behavior and not considering the sequence of action performed by uses. It becomes helpful if we follow user’s session to understand complex user behavior. Therefore to eliminates all these issues this paper proposes a linear-temporal logic model checking approach for the analysis of structured e-commerce weblogs. If we consider a common way of mapping log records according to e-commerce structure, weblogs can be easily converted into event logs by which behavior of the user is captured. After getting users behavior by performing different predefine queries to identify different behavioral patterns that consider the different actions performed by a user during a session.

Social media has developed rapidly, which had already affected modern people's life style. Meanwhile, social media has produced huge amounts of data. In research, we find that there is no explicit definition of user's behavior pattern in social network. Therefore, we focused on the research of that and found people who have two types of behaviors which are inseparably intertwined. The two types that mentioned are functional behaviors and social behaviors. We have defined user's behavior, and regarded user's behavior as a vertical timeline which the object of study is individual. With adding some horizontal lines which stand for social behaviors between vertical timelines, the model linked different people. Experiment's research takes Sina Weibo as an object. For obtaining the social behavior pattern, we use the generalized sequential pattern algorithm (GSP) for data mining. And the pattern is represented by the form of [User Behavior User a#x2192; User Behavior User]*n. According to the result and algorithm accuracy confirmation, the frequent behavioral patterns of a certain person can be obtained. After that we can also predict user's behaviors in the future.

In traditional distributed computing systems a few user types are found having rather ``flat profiles, mainly due to the same administrative domain the users belong to. This is quite different in Computational Grids (CGs) in which several user types should co-exist and make use of resources according to hierarchical nature and the presence of the multiple administrative domains. One implication of the existence of different hierarchical levels in CGs is that it imposes different access and usage policies on resources. In this paper we firstly highlight the most common Grid users types and their relationships and access scenarios in CGs corresponded to old (e.g. performance) and new (e.g security) requirements. Then, we identify and analyze new features arising in users' behavior in Grid scheduling, such as dynamic, selfish, cooperative, trustful, symmetric and asymmetric behaviors. We discuss also how computational economy-based approaches, such as market mechanisms, and computational paradigms, such as Neural Networks, can be used to model user requirements and predict users' behaviors in CGs. As a result of this study we have provided a comprehensive analysis of Grid user scenarios than can serve as a basis for application designers in CGs.

For intelligent interfaces attempting to learn a user's interests, the cost of obtaining labeled training instances is prohibitive because the user must directly label each training instance, and few users are willing to do so. We present an approach that circumvents the need for human-labeled pages. Instead, we learn “surrogate” tasks where the desired output is easily measured, such as the number of hyperlinks clicked on a page or the amount of scrolling performed. Our assumption is that these outputs will highly correlate with the user's interests. In other words, by unobtrusively “observing” the user's behavior we are able to learn functions of value. For example, an intelligent browser could silently observe the user's browsing behavior during the day, then use these training examples to learn such functions and gather, during the middle of the night, pages that are likely to be of interest to the user. Previous work has focused on learning a user profile by passively observing the hyperlinks clicked on and those passed over. We extend this approach by measuring user mouse and scrolling activity in addition to user browsing activity. We present empirical results that demonstrate our agent can accurately predict some easily measured aspects of one's use of his or her browser.

Mining Explainable User Interests from Scalable User Behavior Data

Wireless networks and personalized mobile devices are deeply integrated and embedded in our lives. Such wide adoptions of new technologies will impact user behavior and in turn will affect network performance. It is imperative to characterize the fundamental structure of wireless user behavior in order to model, manage, leverage and design efficient mobile networks. One major challenge in characterizing user behavior stems from the significant size and complexity of user behavioral data. Without summarization and dimension reduction, the sheer amount of data does not provide much useful information. The key contribution of the paper is a novel similarity metric based on a matrix representation of mobility preferences and its decomposition. This method provides an efficient way to reduce important spatiotemporal dynamics in user mobility into a few eigen-behavior vectors. This also facilitates nodes to exchange their mobility summaries and determine their mutual similarity locally. Without any assumption on the properties of user population, we use unsupervised learning (clustering) techniques to classify WLAN users. Such a user grouping scheme based on learned user behavior is crucial for applications relying on the usage context of each mobile device (e.g., participatory sensing, social-relationship-aware message forwarding). In this study, using our systematic TRACE approach, we analyze wireless users' behavioral patterns by extensively mining wireless network logs from two major university campuses to showcase its efficacy. While our findings partly validate intuitive repetitive behavioral trends and user grouping, it is surprising to find the qualitative commonalities and striking consistency of user behavior from the two universities. We discover multimodal user behavior for more than 60 percent of the users, and there are hundreds of distinct groups with unique behavioral patterns in both campuses. The sizes of the major groups follow a power-law distribution.

In this paper, we propose a behavior model for a mobile robot with a projector to provide an appropriate information support based on user's interests and situations. We have been researching Ubiquitous Display (UD) and its behavior model. UD is consisted of a mobile robot and the projector with pan-tilt mechanism, and can show the various information by using its projector on various places, such as floors or walls of corridor. However, since the previous behavior model of the UD did not consider user's situations, UD could not provide information appropriately for each user. Thus, in this paper, we propose a new behavior model for UD that can be adapted to user's interests and situations. UD can understand user's situations by using UD's external sensors such as laser range finders (LRF) and RGBD camera. In addition, to understand user's interests, UD communicates with the user by projecting questions. The user can answer the questions by stepping on or pointing to the projection areas. Based on the user's situations and interests, UD changes its behavior. By using the proposed behavior model, user's situation-aware information support by the UD can be achieved. Preliminary experiments show the effectiveness of the proposed model.

7th International Conference on Affective Computing and Intelligent Interaction, ACII 2017, San Antonio, United States, 23-26 October 2017

Web log analysis can be helpful in gaining information about the usability of the web site, web performance, for marketing purposes, or for development of business intelligence tools in e-commerce systems. User segmentation is one of the problems solved in marketing and e-commerce sphere. Various software was developed to support web analysis. However, most of them provide only information through the tools based on statistics. User behavior and interaction with the web site is usually presented by measurement of click through rates, or by identification and sometimes visualization of popular paths only. User segmentation for further analysis (e.g. campaign analysis in marketing, web recommendation, web usage optimization) is usually allowed with the manual selection (often with variable setting). In this paper is presented the automatic user segmentation (clustering) based on the similar user's behavior on the web site. The user's behavior and behavioral patterns are extracted using process mining techniques; further user segmentation is provided by finding communities with similar behavior through two-step hierarchical clustering.

Personal assistant agents for collaborative design environments

In social networks, link establishment among the users is affected by complex factors. In this paper, we try to investigate the internal and external factors that affect the formation of links and propose a three-level hidden Bayesian link prediction model by integrating the user behavior as well as user relationships to link prediction. First, based on the user multiple interest characteristics, a latent Dirichlet allocation (LDA) traditional text modeling method is applied into user behavior modeling. Taking the advantage of LDA topic model in dealing with the problem of polysemy and synonym, we can mine user latent interest distribution and analyze the effects of internal driving factors. Second, owing to the power-law characteristics of user behavior, LDA is improved by Gaussian weighting. In this way, the negative impact of the interest distribution to the high-frequency users can be reduced and the expression ability of interests can be enhanced. Furthermore, taking the impact of common neighbor dependencies in link establishment, the model can be extended with hidden naive Bayesian algorithm. By quantifying the dependencies between common neighbors, we can analyze the effects of external driving factors and combine internal driving factors to link prediction. Experimental results indicate that the model can not only mine user latent interest distribution but also can improve the performance of link prediction effectively.

WeChat is the largest acquaintance social networking platform in China, which has about 938 million monthly active user accounts. WeChat Moments, known as Friends Circle, serves social networking functions in which users can view information shared by friends. This paper addresses the problem of analyzing the patterns of cascading behavior in WeChat Moments. We obtain 229021 information cascades from WeChat Moments, in which more than 5 million users are involved during 45 days. We analyze these cascades from four aspects to understand the patterns of cascading behavior in WeChat Moments, including the patterns of diffusion structure, temporal dynamic, spatial dynamic and user behavior. In addition, the correlations between these patterns are examined. Our findings contribute to promoting products, predicting and even regulating public opinion.