Changes In User Interests Research Articles

Modeling multi-factor user preferences based on Transformer for next point of interest recommendation

With the rapid development of the mobile network and the gradual popularization of mobile devices, more and more users try to find attractive places to visit through WeChat, Twitter applications. In this trend, personalized next point of interest(POI) recommendation in the Location-based Social Network (LBSN) has become the focus of research and practice. Most existing studies capture user interest changes between different days (i.e. weekend and weekday), however, they ignore seasonal factors in time transition and category factors and thus fail to capture seasonal-level and category-level movement patterns in users’ mobile trajectories. Besides, they neglect the relevance between POIs from all users’ trajectory and fail to generate expressive POI embedding representation without constructing trajectory graph, which will reduce the accuracy of the next POI recommendation. To address the issues above, a next POI recommendation method for modeling Multi-factor User Preferences based on Transformer (MUPT) is developed, which consists of a global POI relationship modeling, a local multi-factor user preference modeling and a prediction module. It first learns the collaborative information of users with similar behavior to generate expressive POI embedding representation. Then it captures the personalized movement patterns of users at the POI, category and time levels based on Transformer mechanism in the local module. Especially the seasonal and other fine-grained information on the time series are learned in the time preference modeling part. The prediction module designed tracks the relationship between multi-level motion pattern representation of user check-in behavior and the next POI accessed by the user, and it finally obtains user’s preference probability for next POIs. An extensive experiment has been conducted on four datasets, and the experimental results analysis demonstrates that our proposed MUPT method is superior to other methods in terms of accuracy(ACC), mean reciprocal rank(MRR) and normalized discounted cumulative gain(NDCG).

A Personalized Collaborative Filtering Recommendation System Based on Bi-Graph Embedding and Causal Reasoning.

The integration of graph embedding technology and collaborative filtering algorithms has shown promise in enhancing the performance of recommendation systems. However, existing integrated recommendation algorithms often suffer from feature bias and lack effectiveness in personalized user recommendation. For instance, users' historical interactions with a certain class of items may inaccurately lead to recommendations of all items within that class, resulting in feature bias. Moreover, accommodating changes in user interests over time poses a significant challenge. This study introduces a novel recommendation model, RCKFM, which addresses these shortcomings by leveraging the CoFM model, TransR graph embedding model, backdoor tuning of causal inference, KL divergence, and the factorization machine model. RCKFM focuses on improving graph embedding technology, adjusting feature bias in embedding models, and achieving personalized recommendations. Specifically, it employs the TransR graph embedding model to handle various relationship types effectively, mitigates feature bias using causal inference techniques, and predicts changes in user interests through KL divergence, thereby enhancing the accuracy of personalized recommendations. Experimental evaluations conducted on publicly available datasets, including "MovieLens-1M" and "Douban dataset" from Kaggle, demonstrate the superior performance of the RCKFM model. The results indicate a significant improvement of between 3.17% and 6.81% in key indicators such as precision, recall, normalized discount cumulative gain, and hit rate in the top-10 recommendation tasks. These findings underscore the efficacy and potential impact of the proposed RCKFM model in advancing recommendation systems.

Heterogeneous Graph Neural Network with Personalized and Adaptive Diversity for News Recommendation

The emergence of online media has facilitated the dissemination of news, but has also introduced the problem of information overload. To address this issue, providing users with accurate and diverse news recommendations has become increasingly important. News possesses rich and heterogeneous content, and the factors that attract users to news reading are varied. Consequently, accurate news recommendation requires modeling of both the heterogeneous content of news and the heterogeneous user-news relationships. Furthermore, users’ news consumption is highly dynamic, which is reflected in the differences in topic concentration among different users and in the real-time changes in user interests. To this end, we propose a Heterogeneous Graph Neural Network with Personalized and Adaptive Diversity for News Recommendation (DivHGNN). DivHGNN first represents the heterogeneous content of news and the heterogeneous user-news relationships as an attributed heterogeneous graph. Then, through a heterogeneous node content adapter, it models the heterogeneous node attributes into aligned and fused node representations. With the proposed attributed heterogeneous graph neural network, DivHGNN integrates the heterogeneous relationships to enhance node representation for accurate news recommendations. We also discuss relation pruning, model deployment, and cold-start issues to further improve model efficiency. In terms of diversity, DivHGNN simultaneously models the variance of nodes through variational representation learning for providing personalized diversity. Additionally, a time-continuous exponentially decaying distribution cache is proposed to model the temporal dynamics of user real-time interests for providing adaptive diversity. Extensive experiments on real-world news datasets demonstrate the effectiveness of the proposed method.

Data-driven smoothing approaches for interest modeling in recommendation systems

In recommendation systems, users often click on some items that are distinct from historically clicked items. This verifies the existence of interest gaps between the historical interests reflected by historical behaviors and current interests. The gaps pose key challenges for interest modeling in recommendation systems, i.e., learn user current interests from historical behaviors. To bridge interest gaps, various solutions have been proposed from a model perspective, and mainly focus on capturing interest diversity and interest evolution process (the changes in user interests over time). Different from existing solutions, our key idea is smoothing the historical interests from a data perspective, so that interest gaps can be smoothed and bridged. To realize this idea, we propose two types of data-driven smoothing approaches based on the principle of collaborative filtering, i.e., click-centric smoothing and user-centric smoothing. The approaches can be easily applied to a number of existing solutions and enhance their performances. To ensure that the smoothed interests completely and accurately cover user current interests (coverage and accuracy), the click-centric smoothing leverages the relevant clicks of a click to smooth the interests on the click, and the user-centric smoothing leverages the relevant clicks of a user click sequence to smooth the interests learned from the click sequence. The “smooth” is to enhance the coverage, and “relevant” is to enhance the accuracy. To search the relevant clicks, we develop an iterative Bayesian inference over a click-flow graph, i.e., a new implementation of collaborative filtering. To further enhance the accuracy, we propose new constrained attention networks to generate high-quality smoothing factors from these relevant clicks. Both offline and online experiments show the data-driven smoothing approaches significantly outperform the baselines. The user-centric smoothing has been deployed on the e-commerce platform of Meituan and achieved 1.51% elevation of Click-through & Conversion Rate and 0.53% elevation of Click-through Rate.

TN-MR: topic-aware neural network-based mobile application recommendation

PurposeWith the increasing number of mobile applications, efficiently recommending mobile applications to users has become a challenging problem. Although existing mobile application recommendation approaches based on user attributes and behaviors have achieved notable effectiveness, they overlook the diffusion patterns and interdependencies of topic-specific mobile applications among user groups. mobile applications among user groups. This paper aims to capture the diffusion patterns and interdependencies of mobile applications among user groups. To achieve this, a topic-aware neural network-based mobile application recommendation method, referred to as TN-MR, is proposed.Design/methodology/approachIn this method, first, the user representations are enhanced by introducing a topic-aware attention layer, which captures both the topic context and the diffusion history context. Second, it exploits a time-decay mechanism to simulate changes in user interest. Multitopic user representations are aggregated by the time decay module to output the user representations of cascading representations under multiple topics. Finally, user scores that are likely to download the mobile application are predicted and ranked.FindingsExperimental comparisons and analyses were conducted on the actual 360App data set, and the results demonstrate that the effectiveness of mobile application recommendations can be significantly improved by using TN-MR.Originality/valueIn this paper, the authors propose a mobile application recommendation method based on topic-aware attention networks. By capturing the diffusion patterns and dependencies of mobile applications, it effectively assists users in selecting their applications of interest from thousands of options, significantly improving the accuracy of mobile application recommendations.

DyCARS: A dynamic context-aware recommendation system.

Dynamic recommendation systems aim to achieve real-time updates and dynamic migration of user interests, primarily utilizing user-item interaction sequences with timestamps to capture the dynamic changes in user interests and item attributes. Recent research has mainly centered on two aspects. First, it involves modeling the dynamic interaction relationships between users and items using dynamic graphs. Second, it focuses on mining their long-term and short-term interaction patterns. This is achieved through the joint learning of static and dynamic embeddings for both users and items. Although most existing methods have achieved some success in modeling the historical interaction sequences between users and items, there is still room for improvement, particularly in terms of modeling the long-term dependency structures of dynamic interaction histories and extracting the most relevant delayed interaction patterns. To address this issue, we proposed a Dynamic Context-Aware Recommendation System for dynamic recommendation. Specifically, our model is built on a dynamic graph and utilizes the static embeddings of recent user-item interactions as dynamic context. Additionally, we constructed a Gated Multi-Layer Perceptron encoder to capture the long-term dependency structure in the dynamic interaction history and extract high-level features. Then, we introduced an Attention Pooling network to learn similarity scores between high-level features in the user-item dynamic interaction history. By calculating bidirectional attention weights, we extracted the most relevant delayed interaction patterns from the historical sequence to predict the dynamic embeddings of users and items. Additionally, we proposed a loss function called the Pairwise Cosine Similarity loss for dynamic recommendation to jointly optimize the static and dynamic embeddings of two types of nodes. Finally, extensive experiments on two real-world datasets, LastFM, and the Global Terrorism Database showed that our model achieves consistent improvements over state-of-the-art baselines.

Exploration of the innovative development path of university civic education based on data mining algorithm

Abstract Addressing the issues of a single form, insufficient targeting, a lack of synergy, and the inability to create a customized collaborative education process in the current college Civic Education curriculum setting, This paper develops a recommendation system for a college Civic Education curriculum based on enhanced collaborative filtering technology. It adopts an enhanced algorithm of cooperative filtering based on the combination, and by having to introduce a gradual forgetting curve relying on the time-sensitive change of user interest, it more effectively resolves. The standard algorithm is outdated, inefficient, and not very adaptable. A similar classification method for college Civics courses is then constructed after data mining is performed to filter out the data noise present in the data set. Via empirical validation, it is discovered that the optimization method obtains an average accuracy of 78.83% when the amount of cycles is 40. When the number of cycles is 40, the recall rate of the optimized algorithm is the highest, and the recall rate is 78%. In terms of recommendation popularity comparison, the optimization algorithm popularity starts to be higher than the traditional algorithm when the number of cycles exceeds 80. Therefore, the system has excellent practicality and resilience, and it operates steadily, which is of positive significance for creating an atmosphere of collaborative and win-win university thinking and political education for teachers and students with various forms and individual innovation.

Integrating users’ long-term and short-term interests with knowledge graph to improve restaurant recommendation

User preferences are typically analyzed by recommendation systems for recommending items. However, in most existing restaurant recommendation methods, recommendation similarity (e.g., reviews, sentiment) is considered only from the perspective of users, whereas restaurant attributes and potential connections between restaurants are ignored. Furthermore, the dining interests of users are diverse and not fixed. For improved recommendations, the changes of user interests should be considered. To address these problems, this study proposed a recommendation approach that combines knowledge graphs and long-term and short-term interests of users. In the approach, user preferences were analyzed from the attributes of restaurant using the semantic structure of knowledge graphs to obtain long-term user representations. Next, we filtered recent historical restaurants based on time series to obtain short-term user representations through the gated recurrent unit. To integrate long- and short-term interests, a novel gated structure was designed to obtain dynamic interest weights. Extensive experiments on a real-world dataset revealed that the proposed approach outperformed baseline methods.

Video Recommendation Method Based on Deep Learning of Group Evaluation Behavior Sequences

Massive video resources satisfy the interests of users on online video platforms but have led to the problem of the “explosion” of video resources. Meanwhile, some problems will also occur such as the sparse data, difficulty in extracting deep features and dynamic changes in user interests in video recommendation. Aiming at the problems, a video recommendation method is proposed based on the deep learning of group evaluation behavior. Using the Word2Vec word vector model, a video is mapped into a high-dimensional feature vector in an evaluation behavior sequence, a video feature vector library is generated, and a feature vector model of the video sequence is established. The convolutional neural networks (CNN), residual networks, and attention mechanisms are integrated to learn the deep connections between video feature vectors and to predict the candidate video sets. The candidate set is expanded by cosine similarity, and a dynamic interest model is established to filter and sort it. Experiments on the Movie-1M dataset show that this method can effectively improve the accuracy and recall rate of video recommendation, which verifies the feasibility and effectiveness of the method.

Tourism Destination Recommendation and Marketing Model Analysis Based on Collaborative Filtering Algorithm

The Internet has penetrated into all fields. As the most dynamic “sunrise industry,” tourism has also been swept into such a wave of Internet. In such an era of “information overload,” how to find one’s favorite attractions among the massive tourist attractions has become a difficult problem. In order to solve this problem, personalized recommendation technology is applied, among which collaborative filtering recommendation technology is one of the core technologies while the collaborative filtering algorithm still has problems. The research and analysis of the algorithm, this paper improves the technology for the problems of low recommendation accuracy that considers user interest changes. It for attribute scoring. It uses the multiattribute score of the item to calculate the user’s overall evaluation score of each attribute of the item; for the change of user interest, a time function based on the Ebbinghaus forgetting law is introduced to calculate the user similarity. It is given a certain weight, that is, a time function, to ultimately ensure the accuracy of the recommendation. Exploring the tourism destination recommendation and marketing model based on the collaborative filtering algorithm can enrich the relevant theories of it on the one hand, and on the other hand, it can lay the foundation for building a real tourism recommendation website.

STRUCTUREAL DEASIGN OF SECURE E-COMMERCE WEBSITES EMPLOYING MULTI-AGENT SYSTEM


 
 
 
 The vast growth in electronic trading and monetary transactions results to increase the importance of E-commerce applications rapidly. In order to accomplish the fully functioning, reliable, scalable, secure, efficient and user-friendly E-commerce applications, sufficient system analysis and design processes are important. Shopping over the internet requires offering secure payment method.
 Providing security for each individual online consumer at the same time (particularly for large websites) is a time-consuming effort that can result in a variety of issues, including response delays, customer orders being lost, and system freeze or crash, all of which impair system performance.
 This work seeks to present a new multi-agent system prototype structure that solves the challenge of security while avoiding issues that might degrade system performance. By creating integrated system taking into account the three parties the e-commerce, the e-bank and the customer, the main concentration is to create an archeticture that helps in decreasing the website traffic and supportng the security with effective performance. To simulate the real systems a recommendation system is created based on the customer behaviors to increase system performance in helping the customers to find their favorite products and solve problems like cold start, sparsity of user‐item matrix, scalability, and changes in user interest.
 This is done by creating a commercial environment (e-commerce and e-bank websites) and a software agent that is installed on the client's device to handle the purchase and ciphering procedures without the need for the consumer to intervene. The security process involve preprocessing phase followed by a modified lightweight AES employing chaotic tent map to generate encryption key.
 
 
 

A collaborative filtering recommendation algorithm based on fuzzy C-means clustering

With the rapid growth of data scale, the problems of collaborative filtering recommendation algorithm are more and more obvious, such as data sparsity, cold start, scalability, and the change of user interest over time. About the existing problems, we introduce the fuzzy clustering and propose a collaborative filtering algorithm based on fuzzy C-means clustering. The algorithm performs fuzzy clustering on the item attribute information to make items belonging to different categories in different membership degree, increases the data density, effectively reduces the data sparsity, and solves the issue that the inaccuracy of similarity leads to the low recommendation accuracy. Meanwhile, the algorithm introduces the time weight function. Different evaluation times give different time weight values, and recently evaluated items are more representative of the user current interest, so we give a higher weight value, and early evaluated items have less effect on the user current interest, thus the weight value are relatively lower. The experimental results show that our algorithm can effectively alleviate the data sparsity problem and time migration of users preferences, thus achieve better performance.

Multi-interaction fusion collaborative filtering for social recommendation

GNNs(Graph Neural Networks) use graph structure to make recommendations, receiving more and more attention. Firstly, existing work focuses on aggregating social interaction information, ignoring users who rate on the same items. Secondly, the existing recommendation methods cannot dynamically reflect changes in user interests. Thirdly, existing methods do not take into account the interaction of subgraphs in GNNs and interactions between user and item factors. In this paper, a graph social fusion recommendation (GSFR) method is proposed. GSFR captures multiple social information simultaneously, based on which it can dynamically adjust user interest weight. Specifically, GSFR captures the interactions in subgraphs with a dynamic attention mechanism, which can represent changes in user interests in heterogeneous networks. A mutualistic mechanism is combined to simulate the mutually reinforcing relationship between social behavior and virtual behavior. User and item latent factors are obtained based on space vectors from the aggregation part. Recommendations are made from inherent characteristics of space vectors interaction behaviors. Comprehensive experimental results on three public datasets show the effectiveness of the proposed model.

Hybrid Recommendation System Based on Collaborative and Content-Based Filtering

A Recommendation System (RS) is a method which filters the information and helped users’ to choose the corresponding target from the huge amount of information obtainable in online. The system recommends useful and satisfactory products (items) such as books, music, jokes, and movies for targeting users based on their interest. The content-based filtering as well as collaborative are different systems used often while designing the RS that predicts the recommended item(s) based upon the user preferences. However, the collaborative filtering algorithm provides poor performance for data sparsity, and it is complex for tracking the change of user interest. Moreover, the hybrid system has combined both the techniques in multiple ways to overcome the shortcomings and optimize the outcomes. Thereby, this article plans to develop a new hybrid recommender system assisting with the optimization concept for optimal recommendation list based on user preference or interest. At first, the feature extraction process takes place, in which the content features and the collaborative features are extracted based on (a) profile construction, (b) content similarity index, (c) Neighbor finder, (d) Items generator, and (e) Items weight generator and variance generator. Consequently, the optimal recommendation is carried out on the basis of features extracted. Further, the developed work plans to carry out the optimal rating of recommendation using a FireFly with Weighted Crow Search Algorithm (FF-WCSA). At last, the outcomes of the developed model is computed to extant approaches in terms of various metrics like accuracy, FDR, MAE, MARE, MSE, MSRE, RMSE, and RMSRE, respectively.

Data Acquisition Method of Sensor News Based on Collaborative Filtering Algorithm

With the vigorous development of new media technologies such as Internet of Things, big data, and cloud computing, data-based sensor news (SN) will become the trend of news reporting in the future and the new normal of news production. Under this background, this paper further analyzes the relationship between SN production mode and traditional news production, including the inheritance of traditional news production value concept, as well as the breakthrough and change in form, media, and effect. In this paper, collaborative filtering (CF) algorithm is improved to solve the problems of data sparseness, user interest migration, and scalability in CF technology. In the calculation of news content similar degree (SD), the influence of part of speech and position of feature words in news is also considered, and the time window is used to establish a model that adapts to the change of user interest with time. In this method, the contribution degree of different attributes to distinguishing users is considered, and the attribute SD between users is accurately calculated, which effectively improves the accuracy of SN data acquisition results.

Modeling the trajectories of interests and preferences of users in digital social systems

Social networks contain a lot of information about users, analyzing which one can draw a conclusion, among other things, about the interests of these users. This paper presents the approaches for static and dynamic analysis of digital footprints of social systems users’, which represents both individual users and user communities in the space of their interests. To do this, topic models were built on such user data as posts and groups, based on the obtained probabilities, multimodal graphs were built, the vertices of which are the received user interests. To analyze changes in user interests, the concept of an interest trajectory was introduced, which allows you to visually monitor the change in the interests of one user, a group of users, and also numerically evaluate this change. The scientific novelty of our approach lies in the proposed method for constructing a multimodal space of user interests and a model of user interest trajectories.

Research on recommendation algorithm of Graph attention Network based on Knowledge graph

The introduction of knowledge graph as the auxiliary information of recommendation system provides a new research idea for personalized intelligent recommendation. However, most of the existing knowledge graph recommendation algorithms fail to effectively solve the problem of unrelated entities, leading to inaccurate prediction of potential preferences of users. To solve this problem, this paper proposes a KG-IGAT model combining knowledge graph and graph attention network, and adds an interest evolution module to graph attention network to capture user interest changes and generate top-N recommendations. Finally, experimental comparison between the proposed model and other algorithms using public data sets shows that KG-IGAT has better recommendation performance.

An E-Commerce Recommendation System Based on Dynamic Analysis of Customer Behavior

The technological development in the devices and services provided via the Internet and the availability of modern devices and their advanced applications, for most people, have led to an increase in the expansion and a trend towards electronic commerce. The large number and variety of goods offered on e-commerce websites sometimes make the customers feel overwhelmed and sometimes make it difficult to find the right product. These factors increase the amount of competition between global commercial sites, which increases the need to work efficiently to increase financial profits. The recommendation systems aim to improve the e-commerce systems performance by facilitating the customers to find the appropriate products according to their preferences. There are lots of recommendation system algorithms that are implemented for this purpose. However, most of these algorithms suffer from several problems, including: cold start, sparsity of user-item matrix, scalability, and changes in user interest. This paper aims to develop a recommendation system to solve the problems mentioned before and to achieve high realistic prediction results this is done by building the system based on the customers’ behavior and cooperating with the statistical analysis to support decision making, to be employed on an e-commerce site and increasing its performance. The project contribution can be shown by the experimental results using precision, recall, F-function, mean absolute error (MAE), and root mean square error (RMSE) metrics, which are used to evaluate system performance. The experimental results showed that using statistical methods improves the decision-making that is employed to increase the accuracy of recommendation lists suggested to the customers.

RESEARCH ON PERSONALIZED RECOMMENDATION ALGORITHM FUSING TIME AND LOCATION

With development of recommendation systems, they are faced with more and more challenges. In order to relieve problems existing in commodity selection by users of different preferences from different regions, personalized recommendation based on location information has emerged. Nowadays most recommendation systems based on location information neglect the fact that users’ preference will change with time. To solve the above problem, geographic location and time factor of users are effectively combined in this paper, and a personalized recommendation algorithm TLPR combining time and location information is proposed. This algorithm determines the users’ geographic location according to postcode information of the users, uses pyramid quadtree model to distribute users into nodes at each layer in the pyramid, utilizes collaborative filtering algorithm for local recommendation in each node, introduces a time function to regulate time-dependent change of user interests when calculating user similarity at each node and finally realizes a comprehensive recommendation by distributing a weight for recommendation result at each layer in the pyramid quadtree. A comparative experience is carried out for recommendation performance of this algorithm on MovieLens dataset, and experimental results indicate that this algorithm is of better recommendation effect

Bidirectional sensing of user preferences and application changes for dynamic mobile app recommendations

Recent years have witnessed the rapid adoption of mobile devices and significant growth in the use of mobile apps. However, the large number of mobile apps makes it difficult for users to determine which ones are of interest. Current app recommendation methods focus on the similarities between user interest and mobile app functionality at a single moment, but they do not consider how these change over time and do not address the underlying mutual effect of such changes. This paper describes a bidirectional sensing method for user interest and app changes that enables precise mobile app recommendations. The proposed method applies continuous bidirectional sensing to both user interest and the functional changes to apps and obtains the evolution trajectory of both. Then, a personal interest evolution network is used to learn the relationship between the two trajectories. Finally, user interest in an app is predicted, and the top-N strategy is used to recommend suitable apps. Experiments on real data sets show that our method can accurately sense user interests and functional changes to apps, and that recommendations based on the personal interest evolution network correspond to actual recommendation scenarios.