Constructing User Profiles Research Articles

Cross-border E-commerce Business Data Processing in the Background of Digital Economy

Abstract This paper analyzes the cross-border e-commerce industry chain, and explores its customer value and impact on marketing strategy through data mining techniques.The whole process data mining methodology of data cleaning, integration, selection, transformation, mining, pattern evaluation and knowledge representation is adopted to deal with the cross-border e-commerce data.This paper focuses on the analysis of the customer value, the construction of the user profile, and the user behavioral patterns based on RFM model.The results show that the customer value is mainly determined by the customer life cycle and the purchase rate. The results show that the customer life cycle and purchase rate mainly determines the customer value of cross-border e-commerce.In the construction of user profiles, the characteristics and behavioral patterns of different users can be effectively identified through data standardization and the construction of user attribute models.The RFM model reveals the different levels of the user’s activity level, consumption ability and consumption value, which provides the basis for formulating differentiated marketing strategies.The data show that among the 4,130 user samples, the user behaviors and behavioral patterns were identified through clustering and knowledge representation. In the sample of 4130 users, cluster analysis shows that most of the users (72.64%) are ordinary customers. In contrast, the proportion of high-value customers (VIP customers) is only 0.24%, which suggests that cross-border e-commerce platforms need to formulate corresponding service and marketing strategies for different user groups to improve customer satisfaction and purchase conversion rate. At the same time, the cross-border e-commerce industry is experiencing changes such as platform fission, marketing innovation, and sinking of service targets, which poses new challenges and opportunities for e-commerce platforms’ operation and management.

Big Data-Driven Cross-Border E-commerce Platform Operation Strategy Based on Data Mining

Abstract In the context of the big data era, cross-border e-commerce enterprises are confronting significant challenges. The traditional marketing model finds it challenging to adapt to the evolving needs, making precision marketing for cross-border e-commerce platforms particularly crucial. This paper, based on the big data-driven path of operating cross-border e-commerce platforms, designs the operation strategy of these platforms from the perspective of precision marketing and empirically analyzes its impact. The RFM model is first used to design user value labels, and the K-means algorithm then uses the clustered labels. Combined with mining the three types of key data—user basic attributes, user value labels, and user consumption behaviors—of Company A’s Amazon store to construct user profiles, analyze them, and further design precise marketing strategies based on user profiles and analyze their effects, This paper classifies customer groups into three categories: high-value premium types, dynamic premium types, and growth types. High-value, quality customers account for most of Company A’s business, and through personalized marketing, their sales show a certain growth trend. Vitality-quality customers stimulate the desire to buy by recommending new products and activating old ones, and their sales increased to $32,527 in the fourth quarter. The impact of growth-type customers using consumption coupons and discount codes to stimulate consumption and purchases is flat, with no significant growth. This indicates that the operation strategy in this paper is more obvious and can be used as a precursor for further optimization.

Research on User Profile Combined with Collaborative Filtering Recommendation Algorithm for Intelligent Tourism

In recent years, the online travel sector in the tourism industry has experienced significant growth and popularity due to the development and widespread adoption of internet technology and smart devices. However, despite these advancements, scenic spots have struggled to provide precise services to tourists, as the online marketplace is flooded with numerous and disorganized commodity resources, lacking standardized construction and systematic management. As a result, travelers find it challenging to access specialty goods that cater to their personalized needs. To address this issue, this paper proposes the use of user profiling and collaborative filtering recommendation algorithms to achieve personalized recommendations for specialty products in scenic spots. The general process of constructing user profiles for scenic spots and combining them with collaborative filtering algorithms to create an intelligent tourism recommendation system is outlined. The paper also highlights the current challenges faced by this system in practical applications and provides future research prospects to promote accurate services in tourist attractions.

Influence Maximization in Trajectory Database

Abstract: In this project, we study a novel problem of influence maximization in trajectory databases that is very useful in precise location-aware advertising. It finds k best trajectories to be attached with a given advertisement and maximizes the expected influence among a large group of audience. We show that the problem is NP-hard and propose both exact and approximate solutions to find the best set of trajectories. In the exact solution, we devise an expansion-based framework that enumerates trajectory combinations in a best-first manner and propose three types of upper bound estimation techniques to facilitate early termination. In addition, we propose a novel trajectory index to reduce the influence calculation cost. To support large k, we propose a greedy solution with an approximation ratio of (1-1/e), whose performance is further optimized by a new proposed cluster-based method. We also propose a threshold method that can support any approximation ratio ∈ (0, 1]. In addition, we extend our problem to support the scenario when there are a group of advertisements. In our experiments, we use real datasets to construct user profiles, motion patterns and trajectory databases. The experimental results verified the efficiency of our proposed methods.

EXPERT RECOMMENDATION THROUGH TAG RELATIONSHIP IN COMMUNITY QUESTION ANSWERING

Community Question Answering (CQA) services are technical discussion forums websites on social media that serve as a platform for users to interact mainly via question and answer. However, users of this platform have posed dissatisfaction over the slow response and the preference for user domains due to the overwhelming information in CQA websites. Numerous past studies focusing on expert recommendation are solely based on the information available from websites where they rarely account for the preference of users’ domain knowledge. This condition prompts the need to identify experts for the questions posted on community-based websites. Thus, this study attempts to identify ranking experts’ derived from the tag relationship among users in the CQA websites to construct user profiles where their interests are realized in the form of tags. Experts are considered users who post high-quality answers and are often recommended by the system based on their previous posts and associated tags. These associations further describe tags that often co-occur in posts and the significant domains of user interest. The current study further explores this relationship by adopting the “Tag Relationship Expert Recommendation (TRER)” method where Questions Answer (QA) Space is utilized as a dataset to identify users with similar interests and subsequently rank experts based on the tag-tag relationship for user’s question. The results show that the TRER method outperforms existing baseline methods by effectively improving the performance of relevant domain experts in CQA, thereby facilitating the expert recommendation process in answering questions posted by technical and academic professionals.

Fed-UserPro: A user profile construction method based on federated learning

Abstract User profiles constructed using vast network behaviour data are widely used in various fields. However, data island and central server capacity problems limit the implementation of centralised big data training. This paper proposes a user profile construction method, Fed-UserPro, based on federated learning, which uses non-independent and identically distributed unstructured user text to jointly construct user profiles. Latent Dirichlet allocation model and softmax multi-classification regression method are introduced into the federated learning structure to train data. The results show that the accuracy of the Fed-UserPro method is 8.69%–19.71% higher than that of single-party machine learning methods.

Hierarchical reinforcement learning with dynamic recurrent mechanism for course recommendation

In online learning scenarios, the learners usually hope to find courses that meet their preferences and the needs for their future developments. Thus, there is a great need to develop effective personalized course recommender systems that can guide the learners to choose suitable courses. In practice, Reinforcement Learning (RL) can be applied to build dynamic user profiles from users’ interactions with courses, which is the key to the success of a course recommender system. However, existing RL-based course recommendation methods usually suffer from the trade-off between exploration and exploitation. In this paper, we propose a novel course recommendation model, namely Hierarchical rEinforcement Learning with dynAmic Recurrent mechanism (HELAR), in which a profile constructor with autonomous learning ability is designed to make personalized course recommendation. To address the exploration-exploitation trade-off issue in constructing user profiles, we propose a novel policy gradient method. It employs a recurrent scheme by context-aware learning to exploit the current knowledge, while utilizing a dynamic baseline to explore the user’s future preferences. Extensive experiments are conducted on two real-world datasets to evaluate the performance of the proposed HELAR model, and the experimental results demonstrate the advantage of HELAR over state-of-the-art course recommendation methods.

User Demographic Prediction Based on the Fusion of Mobile and Survey Data

The user demographic prediction problem is one of the critical processes in the construction of user profiles, which is of great significance for understanding users’ characteristics and attributes. Most of the prior works on this problem either used only single-source data or employed a hard-matching method to handle multi-source data. These methods will result in a great loss of data and information in many circumstances, which may affect the model’s accuracy as well as the application scenarios. In order to solve these problems, this paper proposes a framework for user demographic prediction based on mobile and survey data, and presents a Deep Structured Fusion Model (DSFM) using neural networks with attention mechanisms to perform data fusion by comparing user similarity between two heterogeneous datasets. We examine the effectiveness of the framework and the fusion model on a real-world mobile dataset with almost one billion users, using a survey dataset containing 29,809 users’ questionnaire results as an additional information source to predict users’ age and gender. Our framework achieves excellent results on these datasets, increasing the prediction accuracy of gender and age by up to 3.23% and 5.21% compared to the best baseline model.

Who Am I?

To enable targeted ads, companies profile Internet users, automatically inferring potential interests and demographics. While current profiling centers on users' web browsing data, smartphones and other devices with rich sensing capabilities portend profiling techniques that draw on methods from ubiquitous computing. Unfortunately, even existing profiling and ad-targeting practices remain opaque to users, engendering distrust, resignation, and privacy concerns. We hypothesized that making profiling visible at the time and place it occurs might help users better understand and engage with automatically constructed profiles. To this end, we built a technology probe that surfaces the incremental construction of user profiles from both web browsing and activities in the physical world. The probe explores transparency and control of profile construction in real time. We conducted a two-week field deployment of this probe with 25 participants. We found that increasing the visibility of profiling helped participants anticipate how certain actions can trigger specific ads. Participants' desired engagement with their profile differed in part based on their overall attitudes toward ads. Furthermore, participants expected algorithms would automatically determine when an inference was inaccurate, no longer relevant, or off-limits. Current techniques typically do not do this. Overall, our findings suggest that leveraging opportunistic moments within pervasive computing to engage users with their own inferred profiles can create more trustworthy and positive experiences with targeted ads.

Automatic construction of multi-faceted user profiles using text clustering and its application to expert recommendation and filtering problems

In the information age we are living in today, not only are we interested in accessing multimedia objects such as documents, videos, etc. but also in searching for professional experts, people or celebrities, possibly for professional needs or just for fun. Information access systems need to be able to extract and exploit various sources of information (usually in text format) about such individuals, and to represent them in a suitable way usually in the form of a profile. In this article, we tackle the problems of profile-based expert recommendation and document filtering from a machine learning perspective by clustering expert textual sources to build profiles and capture the different hidden topics in which the experts are interested. The experts will then be represented by means of multi-faceted profiles. Our experiments show that this is a valid technique to improve the performance of expert finding and document filtering.

HIBoosting: A Recommender System Based on a Gradient Boosting Machine

Based on explicit data, collaborative filtering is one of the most valuable technologies of a recommender system. However, the further development of a recommender system has been restricted to some extent by the problems of cold start and data sparsity. To weaken the effect caused by data loss, some implicit feedback data are typically introduced into the recommendation such as social information and attribute data. In this paper, we propose a heterogeneous information boosting model for recommendations (HIBoosting). The model takes advantage of different network meta-paths to mine the potential information of user social networks, object metadata and interactive data in heterogeneous information networks. For different semantic information, we use a random-walk based on a meta-path to calculate the similarities of users or items, which are used to extract the low-dimensional embedded features of users or items by factorization, construct user profiles and item profiles. Finally, the embedded features are used for integrated learning by a gradient boosting decision tree (GBDT) to provide users with better recommendation services. The model not only integrates heterogeneous information in information networks, but also makes full use of mining latent relations of heterogeneous information networks. Experiments on the CiaoDVD and Lastfm datasets show that HIBoosting has made great progress in the accuracy of a recommender system.

Rating prediction for recommendation: Constructing user profiles and item characteristics using backpropagation

Recommender systems play an important role in quickly identifying and recommending most acceptable products to the users. The latent user factors and item characteristics determine the degree of user satisfaction on an item. While many of the methods in the literature have assumed that these factors are linear, there are some other methods that treat these factors as nonlinear; but they do it in a more implicit way. In this paper, we have investigated the effect of true nature (i.e., nonlinearity) of the user factors and item characteristics, and their complex layered relationship on rating prediction. We propose a new deep feedforward network that learns both the factors and their complex relationship concurrently. The aim of our study was to automate the construction of user profiles and item characteristics without using any demographic information and then use these constructed features to predict the degree of acceptability of an item to a user. We constructed the user and item factors by using separate learner weights at the lower layers, and modeled their complex relationship in the upper layers. The construction of the user profiles and the item characteristics, solely based on rating triples (i.e., user id, item id, rating), overcomes the requirement of explicit demographic information be given to the system. We have tested our model on three real world datasets: Jester, Movielens, and Yahoo music. Our model produces better rating predictions than some of the state-of-the-art methods which use demographic information. The root mean squared error incurred by our model on these datasets are 4.0873, 0.8110, and 0.9408 respectively. The errors are smaller than current best existing models’ errors in these datasets. The results show that our system can be integrated to any web store where development of hand engineered features for recommending products is less feasible due to huge traffics and also that there is a lack of demographic information about the users and the items.

Location-Sensitive User Profiling Using Crowdsourced Labels

In this paper, we investigate the impact of spatial variation on the construction of location-sensitive user profiles. We demonstrate evidence of spatial variation over a collection of Twitter Lists, wherein we find that crowdsourced labels are constrained by distance. For example, that energy in San Francisco is more associated with the green movement, whereas in Houston it is more associated with oil and gas. We propose a three-step framework for location-sensitive user profiling: first, it constructs a crowdsourced label similarity graph, where each labeler and labelee are annotated with a geographic coordinate; second, it transforms this similarity graph into a directed weighted tree that imposes a hierarchical structure over these labels; third, it embeds this location-sensitive folksonomy into a user profile ranking algorithm that outputs a ranked list of candidate labels for a partially observed user profile. Through extensive experiments over a Twitter list dataset, we demonstrate the effectiveness of this location-sensitive user profiling.

Seed Selection for Data Offloading Based on Social and Interest Graphs

The explosive growth of mobile data demand is becoming an increasing burden on current cellular network. To address this issue, we propose a solution of opportunistic data offloading for alleviating overloaded cellular traffic. The principle behind it is to select a few important users as seeds for data sharing. The three critical steps are detailed as follows. We first explore individual interests of users by the construction of user profiles, on which an interest graph is built by Gaussian graphical modeling. We then apply the extreme value theory to threshold the encounter duration of user pairs. So, a contact graph is generated to indicate the social relationships of users. Moreover, a contact-interest graph is developed on the basis of the social ties and individual interests of users. Corresponding on different graphs, three strategies are finally proposed for seed selection in an aim to maximize overloaded cellular data. We evaluate the performance of our algorithms by the trace data of real-word mobility. It demonstrates the effectiveness of the strategy of taking social relationships and individual interests into account.

A Dynamic Personalized News Recommendation System Based on BAP User Profiling Method

In this paper, we propose a user profile model to describe users’ preferences from multiple perspectives. Then, we discuss the degree of the user’s preferences for historical news, and propose a method to calculate the preference weight of historical news according to the user’s reading behavior and the popularity of news. This method could construct user profiles more accurately. Besides, we provide a dynamic method for news recommendation, in which both short-term and long-term user preferences are considered. The experimental results indicate that our method can significantly improve the recommendation effect.

Predicting users’ demographic characteristics in a Chinese social media network

PurposeBased on user-generated content from a Chinese social media platform, this paper aims to investigate multiple methods of constructing user profiles and their effectiveness in predicting their gender, age and geographic location.Design/methodology/approachThis investigation collected 331,634 posts from 4,440 users of Sina Weibo. The data were divided into two parts, for training and testing . First, a vector space model and topic models were applied to construct user profiles. A classification model was then learned by a support vector machine according to the training data set. Finally, we used the classification model to predict users’ gender, age and geographic location in the testing data set.FindingsThe results revealed that in constructing user profiles, latent semantic analysis performed better on the task of predicting gender and age. By contrast, the method based on a traditional vector space model worked better in making predictions regarding the geographic location. In the process of applying a topic model to construct user profiles, the authors found that different prediction tasks should use different numbers of topics.Originality/valueThis study explores different user profile construction methods to predict Chinese social media network users’ gender, age and geographic location. The results of this paper will help to improve the quality of personal information gathered from social media platforms, and thereby improve personalized recommendation systems and personalized marketing.

Influence Maximization in Trajectory Databases

In this paper, we study a novel problem of influence maximization in trajectory databases that is very useful in precise location-aware advertising. It finds $k$ best trajectories to be attached with a given advertisement and maximizes the expected influence among a large group of audience. We show that the problem is NP-hard and propose both exact and approximate solutions to find the best set of trajectories. In the exact solution, we devise an expansion-based framework that enumerates trajectory combinations in a best-first manner and propose three types of upper bound estimation techniques to facilitate early termination. In addition, we propose a novel trajectory index to reduce the influence calculation cost. To support large $k$ , we propose a greedy solution with an approximation ratio of (1 − 1/e), whose performance is further optimized by a new proposed cluster-based method. We also propose a threshold method that can support any approximation ratio $\epsilon \in (0,1]$ . In addition, we extend our problem to support the scenario when there are a group of advertisements. In our experiments, we use real datasets to construct user profiles, motion patterns, and trajectory databases. The experimental results verified the efficiency of our proposed methods.

Exploiting search history of users for news personalization

Content personalization is a long-standing problem for online news services. In most personalization approaches users of a news service are represented by topical interest profiles that are matched with news articles in order to properly decide which articles are to be recommended. When constructing user profiles, existing personalization methods exploit the user activity observed within the news service itself without incorporating additional information that can be obtained from other sources.In this paper we study the problem of news personalization by leveraging usage information that is external to the news service. We propose a novel approach that relies on the concept of “search profiles”, which are user profiles that are built based on the past interactions of the user with a web search engine. We extensively test our proposal on real-world datasets obtained from Yahoo. We explore various dimensions and granularities at which search profiles can be built. Experimental results show that, compared to a basic strategy that does not exploit the search activity of users, our approach is able to boost the clicks on news articles shown at the top positions of a ranked result list.

Intelligence Based User Profile Generation

Web intelligence provides a platform that empowers internet users to determine the most appropriate and best information for their interests. It provides the ability to sense and adapt to the needs and preference of the user. The recent advancements have made it conceivable to capture the users experience and interactions with web. Consequently predicting users behaviors will expedite and enhance browsing experience. This paper proposes an intelligent approach for making the web more powerful by predicting the conduct of individual users. The main goal is to implicitly construct user profiles using a Particle Swarm Optimization - based technique. We reveal interesting results in comparing with a standard user modeling approach.

Community-aware user profile enrichment in folksonomy

In the era of big data, collaborative tagging (a.k.a. folksonomy) systems have proliferated as a consequence of the growth of Web 2.0 communities. Constructing user profiles from folksonomy systems is useful for many applications such as personalized search and recommender systems. The identification of latent user communities is one way to better understand and meet user needs. The behavior of users is highly influenced by the behavior of their neighbors or community members, and this can be utilized in constructing user profiles. However, conventional user profiling techniques often encounter data sparsity problems as data from a single user is insufficient to build a powerful profile. Hence, in this paper we propose a method of enriching user profiles based on latent user communities in folksonomy data. Specifically, the proposed approach contains four sub-processes: (i) tag-based user profiles are extracted from a folksonomy tripartite graph; (ii) a multi-faceted folksonomy graph is constructed by integrating tag and image affinity subgraphs with the folksonomy tripartite graph; (iii) random walk distance is used to unify various relationships and measure user similarities; (iv) a novel prototype-based clustering method based on user similarities is used to identify user communities, which are further used to enrich the extracted user profiles. To evaluate the proposed method, we conducted experiments using a public dataset, the results of which show that our approach outperforms previous ones in user profile enrichment.