Recommendation In Social Networks Research Articles

An approach to the mobile social services recommendation algorithm based on association rules mining

With the continuous development of social network, more and more research areas have emerged. Among them, service recommendation has always been a research hotspot. Yet achieving accurate personalised recommendation is challenging because of the large amount of data generated in social networks as the data available for each user is highly sparse. In this paper, it focuses on the association rules in the mobile social network recommendation algorithms. By introducing the mobile users' location information to the collaborative filtering recommendation process, the association rules between the items are minded. Then the association rules are filtered and split, which are integrated into the similarity matrix to combine user location information with them. Also, an association rule mining model is proposed, which considers time factor to further improved the recommendation accuracy. Experimental results show that the performance of our algorithm is better than the baseline.

A fine-grained social network recommender system

Recommender systems have greatly evolved in recent years and have become an integral part of the Web. From e-commerce sites to mobile apps, our daily routine revolves around a series of “small” decisions that are influenced by such recommendations. In a similar manner, online social networks recommend only a subset of the massive amount of content published by a user’s friends. However, the prevalent approach for the content selection process in such systems is driven by the amount of interaction between the user and the friend who published the content. As a result, content of interest is often lost due to weak social ties. In this paper, we present a fine-grained recommender system for social ecosystems, designed to recommend media content (e.g., music videos, online clips) published by the user’s friends. The system design was driven by the findings of our qualitative user study that explored the value and requirements of a recommendation component within a social network. The core idea behind the proposed approach was to leverage the abundance of preexisting information in each user’s account for creating interest profiles, to calculate similarity scores at a fine-grained level for each friend. The intuition behind the proposed method was to find consistent ways to obtain information representations that can identify overlapping interests in very specific sub-categories (e.g., two users’ music preferences may only coincide on hard rock). While the system is intended as a component of the social networking service, we developed a proof-of-concept implementation for Facebook and explored the effectiveness of our underlying mechanisms for content analysis. Our experimental evaluation demonstrates the effectiveness of our approach, as the recommended content of interest was both overlooked by the existing Facebook engine and not contained in the users’ Facebook News Feed. We also conducted a user study for exploring the usability aspects of the prototype and found that it offers functionality that could significantly improve user experience in popular services.

A topic sentiment based method for friend recommendation in online social networks via matrix factorization

Data sparsity and prediction quality have been recognized as the crucial challenges in recommender system. With the expansion of social network data, social network analysis is becoming more and more important. Traditional Recommendation System assumes that users are independent and distributed equally, which ignores the social interaction or connection among users. In order to solve the prediction quality of friend recommendation in social networks, a user recommendation algorithm for social networks based on sentiment analysis and matrix factorization is proposed in this paper. This method is based on the traditional matrix factorization model. By integrating Sentiment (S), Important (I) and Objective (O) of user topic content in the social network, this paper proposes the approach base on sentiment analysis and matrix factorization to solve the poor prediction accuracy by employing social network. SIO model solves the problem that users in social networks can′t score the content of topics. User-topic matrix is constructed by SIO model. Combining the SIO model with matrix factorization, algorithm called SIO-TMF algorithm is proposed. Applying this method on social network, comparing with some traditional recommendation algorithms from four aspects: accuracy, diversity, novelty and coverage, the experimental results show that the proposed method improves the prediction quality of recommender system.

A Matrix Factorization Algorithm for Efficient Recommendations in Social Rating Networks Using Constrained Optimization

In recent years the emergence of social media has become more prominent than ever. Social networking has become the de facto tool used by people all around the world for information discovery. Consequently, the importance of recommendations in a social network setting has urgently emerged, but unfortunately, many methods that have been proposed in order to provide recommendations in social networks cannot produce scalable solutions, and in many cases are complex and difficult to replicate unless the source code of their implementation has been made publicly available. However, as the user base of social networks continues to grow, the demand for developing more efficient social network-based recommendation approaches will continue to grow as well. In this paper, following proven optimization techniques from the domain of machine learning with constrained optimization, and modifying them accordingly in order to take into account the social network information, we propose a matrix factorization algorithm that improves on previously proposed related approaches in terms of convergence speed, recommendation accuracy and performance on cold start users. The proposed algorithm can be implemented easily, and thus used more frequently in social recommendation setups. Our claims are validated by experiments on two real life data sets, the public domain Epinions.com dataset and a much larger dataset crawled from Flixster.com.

Friend recommendation for cross marketing in online brand community based on intelligent attention allocation link prediction algorithm

Circle structure of online brand communities allows companies to conduct cross-marketing activities by the influence of friends in different circles and build strong and lasting relationships with customers. However, existing works on the friend recommendation in social network do not consider establishing friendships between users in different circles, which has the problems of network sparsity, neither do they study the adaptive generation of appropriate link prediction algorithms for different circle features. In order to fill the gaps in previous works, the intelligent attention allocation link prediction algorithm is proposed to adaptively build attention allocation index (AAI) according to the sparseness of the network and predict the possible friendships between users in different circles. The AAI reflects the amount of attention allocated to the user pair by their common friend in the triadic closure structure, which is decided by the friend count of the common friend. Specifically, for the purpose of overcoming the problem of network sparsity, the AAIs of both the direct common friends and indirect ones are developed. Next, the decision tree (DT) method is constructed to adaptively select the suitable AAIs for the circle structure based on the density of common friends and the dispersion level of common friends’ attention. In addition, for the sake of further improving the accuracy of the selected AAI, its complementary AAIs are identified with support vector machine model according to their similarity in value, direction, and ranking. Finally, the mutually complementary indices are combined into a composite one to comprehensively portray the attention distribution of common friends of users in different circles and predict their possible friendships for cross-marketing activities. Experimental results on Twitter and Google+ show that the model has highly reliable prediction performance.

Discovering Memory-Based Preferences for POI Recommendation in Location-Based Social Networks

Point-of-interest (POI) recommendations in location-based social networks (LBSNs) allow online users to discover various POIs for social activities occurring in the near future close to their current locations. Research has verified that people’s preferences regarding POIs are significantly affected by various internal and external contextual factors, which are therefore worth extensive study for POI recommendation. However, although psychological effects have also been demonstrated to be significantly correlated with an individual’s preferences, such effects have been largely ignored in previous studies on POI recommendation. For this paper, inspired by the famous memory theory in psychology, we were interested in whether memory-based preferences could be derived from users’ check-in data. Furthermore, we investigated how to incorporate these memory-based preferences into an effective POI recommendation scheme. Consequently, we refer to Ebbinghaus’s theory on memory, which describes the attenuation of an individual’s memory in the form of a forgetting curve over time. We first created a memory-based POI preference attenuation model and then adopted it to evaluate individuals’ check-ins. Next, we employed the memory-based values of check-ins to calculate the POI preference similarity between users in an LBSN. Finally, based on this memory-based preference similarity, we developed a novel POI recommendation method. We experimentally evaluated the proposed method on a real LBSN data set crawled from Foursquare. The results demonstrate that our method, which incorporates the proposed memory-based preference similarity for POI recommendation, significantly outperforms other methods. In addition, we found the best value of the parameter H in the memory-based preference model that optimizes the recommendation performance. This value of H implies that an individual’s memory usually has an effect on their daily travel choices for approximately 300 days.

PRINCIPLES OF SEARCHING AND SORTING OPTIMIZATION IN SOCIAL NETWORKS USING A MULTI-FACTOR ASSESSMENT SYSTEM

The analysis of social networks, which focuses on the relationship between social entities today is an area of active research. It is a set of tools for research, in particular, in combination with artificial intelligence methods such as machine learning, deep learning. The paper examined the current quality of the assessment of information in social networks, analyzed the methods of searching and sorting information in various social networks, as well as the process of providing recommendations to users. Social media data is an inexhaustible source of research and business opportunities. In general, social media data is information gathered from social networks that shows how users interact with content. Methods of improving search results for personalizing recommendations in social networks are given. These indicators and statistics provide an effective understanding of the strategy of behavior in social networks. The advantages and disadvantages of a multifactor assessment system are considered. The possible ways of integrating the combined system of evaluating information elements by the user to optimize search queries and filtering big data are identified.

Augmented label propagation for seed set expansion

In many applications such as social network analysis and recommendation systems, it is of particular interest to identify a group of similar nodes/users/items. However, in networks of massive size, manual labeling process becomes intractable. A practical means is to mark a small number of nodes as seeds, and then expand them to the rest (unlabeled) ones, which is also known as seed set expansion. We present a novel method for seed set expansion by leveraging information spreading dynamics through label propagation. In particular, by devising an augmented, community-based label propagation, we can fully exploit the information of the limited seed nodes, and apply the connectivity structure of the whole network in imposing a larger number of constraints on the label propagation process, thus achieving an improved estimation. Our method can increase the effective number of seed nodes in that it can achieve a better estimation than other propagation methods using the same number of seeds. Extensive experiments on real-world datasets demonstrate the effectiveness and adaptiveness of our method, compared to the state-of-the-art approaches.

A novel trust recommendation model for mobile social network based on user motivation

Traditional collaborative filtering recommendation algorithm has the problems of sparse data and limited user preference information. To deal with data sparseness problem and the unreliability phenomenon on the traditional social network recommendation. This paper presents a novel algorithm based on trust relationship reconstruction and social network delivery. This paper introduces the method of eliminating falsehood and storing truth to avoid the unreliable phenomenon and improves the accuracy of falsehood according to the user similarity formula based on the scale of contact established by users. In this paper, the problem of attack caused by the misbehaving nodes is investigated when the recommended information is disseminated in the existing trust model. In addition, a recommendation-based trust model is proposed that includes a defensive plan. This scheme employs the clustering techniques on the basis of interaction count, information Compatibility and node intimacy, in a certain period of time dynamically filter dishonest recommendation related attacks. The model has been verified in different portable and detached topologies. The network knots undergo modifications regarding their neighbors as well as frequent routes. The experimental analysis indicates correctness and robustness of the reliance system in an active MANET setting. Compared with the most advanced recommender system, the proposed recommendation algorithm in accuracy and coverage measurements show a significant improvement.

Evolving graph construction for successive recommendation in event-based social networks

Personalized recommendation can help individual users to quickly reserve their interested events, which makes it indispensable in event-based social networks (EBSNs). However, as each EBSN is often with large amount of entities and each upcoming event is normally with non-repetitive uniqueness, how to deal with such challenges is crucial to the success of event recommendation. In this paper, we propose an evolving graph-based successive recommendation (EGSR) algorithm to address such challenges: The basic idea is to exploit the random walk with restart (RWR) on a recommendation graph for ranking the upcoming events. In EGSR, we employ a sliding window mechanism to construct evolving graphs for successively recommending new events for each user. We propose a graph entropy-based contribution measure for adjusting the window length and for weighting the history information. In EGSR, we also apply a topic analysis technique for analyzing event text description. We then propose to establish each user an interest model and to compute the similarities in between event content and user interest as edges’ weights for each recommendation graph. In successive recommendation, the number of upcoming events may experience great variations in different times. For a fair comparison, we also propose a set of cumulative evaluation metrics based on the traditional recommendation performance metrics. Experiments have been conducted based on the crawled one year data from a real EBSN for two cities. Results have validated the superiority of the proposed EGSR algorithm over the peer ones in terms of better recommendation performance and reduced computation complexity.

Intelligent recommendation algorithm for social networks based on gradient particle swarm optimization

Aiming to improve the intelligent recommendation ability of users in social networks, an intelligent recommendation algorithm for social networks based on gradient particle swarm optimization (GPSO) is proposed. Under the limited sample training, the user association relationship model of social network is constructed, and the hybrid kernel function and the global kernel function are constructed to extract the correlation amount feature of the social network recommendation information. The hybrid particle swarm optimization algorithm is used for adaptive learning of social network recommendation, Logistic chaotic mapping is used to control the convergence of recommendation process, and the potential characteristics of network users and the universality and ergodicity of social features are analyzed. The particle swarm optimization and adaptive optimization of social network recommendation are implemented by using gradient algorithm to realize swarm intelligence recommendation of social network user information. The simulation results show that the proposed algorithm is more accurate and convergent, and it avoids falling into the local optimal solution in the process of intelligent recommendation of social network. The intelligence and global stability of social network recommendation are improved.

Location perspective-based neighborhood-aware POI recommendation in location-based social networks

As an effective way to help users find attractive locations and meet their individual needs, point-of-interest (POI) recommendation has become an important application in location-based social networks (LBSNs). Although the geographical influence has been reported as an effective factor for improving POI recommendation accuracy, previous work mainly models it from the user perspective instead of location perspective. Intuitively, neighboring POIs tend to be visited by similar users, which implies that modeling geographical relationships from a location perspective can simulate users’ behavior more reasonably. Moreover, different from traditional recommendation problems, users in LBSNs often express their interests only by checking in different POIs, which is a kind of implicit feedback. In other words, we can easily get the POIs that the users have visited, but it is hard to get the ones that the users do not like. We cannot use a common approach to distinguish these negative values directly. Based on the above observations, this work concentrates on exploiting the geographical relationships among POIs from a location perspective for implicit problem, where a location neighborhood-aware weighted probabilistic matrix factorization is proposed (L-WMF). To be specific, the weighted probabilistic matrix factorization (WMF) that can deal with implicit feedback is first introduced as our basic POI recommendation method. Then, we incorporate the geographical relationships among POIs into the WMF as the regularization terms to exploit the geographical characteristics from a location perspective. Finally, we conduct several experiments to evaluate our L-WMF method on two real-world datasets, and the experimental results indicate that the L-WMF is more effective and can reach better performance than other related methods.

A Joint Deep Recommendation Framework for Location‐Based Social Networks

Location‐based social networks, such as Yelp and Tripadvisor, which allow users to share experiences about visited locations with their friends, have gained increasing popularity in recent years. However, as more locations become available, the need for accurate systems able to present personalized suggestions arises. By providing such service, point‐of‐interest recommender systems have attracted much interest from different societies, leading to improved methods and techniques. Deep learning provides an exciting opportunity to further enhance these systems, by utilizing additional data to understand users’ preferences better. In this work we propose Textual and Contextual Embedding-based Neural Recommender (TCENR), a deep framework that employs contextual data, such as users’ social networks and locations’ geo‐spatial data, along with textual reviews. To make best use of these inputs, we utilize multiple types of deep neural networks that are best suited for each type of data. TCENR adopts the popular multilayer perceptrons to analyze historical activities in the system, while the learning of textual reviews is achieved using two variations of the suggested framework. One is based on convolutional neural networks to extract meaningful data from textual reviews, and the other employs recurrent neural networks. Our proposed network is evaluated over the Yelp dataset and found to outperform multiple state‐of‐the‐art baselines in terms of accuracy, mean squared error, precision, and recall. In addition, we provide further insight into the design selections and hyperparameters of our recommender system, hoping to shed light on the benefit of deep learning for location‐based social network recommendation.

An intelligent recommender system using social trust path for recommendations in web-based social networks

In this paper, we combine a social regularization approach that incorporates social network information to benefit recommender systems with the trust information between users. Both trust and rating records (tags) are employed to predict the missing values (tags) in the user-item matrix. Especially, we use an algorithm for best recommended trust path selection, to identify multiple recommended trust paths and to determine an aggregate path for generating different final recommendations. Empirical analyses on real datasets show that the combination of social information and trust achieves superior performance to existing approaches.

Group Event Recommendation in Event-Based Social Networks Considering Unexperienced Events

As users in event-based social networks (EBSNs) usually participate in events together with friends, classmates, colleagues, or family members, recommending events to a group of users is attracting increasing attention in EBSNs recent years. However, existing studies are lack of attention to the fact that groups in EBSNs may have potential desires for participating in the unexperienced events. In order to deal with the challenges, including mining implicit friendships between users, simulating the consulting process between users and their friends outside the groups, and simulating the negotiating process among members inside the groups, we propose a t wo- p hase g roup e vent r ecommendation (2PGER) model for EBSNs. First, we leverage information, such as online social behaviors, users’ event participation records, and topological structures of EBSNs to establish a global trust network among users and establish egotrust networks of all users. Then, we perform random walks on the pre-built egotrust network for each user to acquire the user’s predicted preferences on the unexperienced events. Third, we adopt a random walk with restarts (RWR) method to aggregate users’ preferences and recommend top N events to groups. In the end, we compare 2PGER with several baseline approaches on real datasets from Meetup. The results show that 2PGER outperforms baselines.

Semantic-Enhanced and Context-Aware Hybrid Collaborative Filtering for Event Recommendation in Event-Based Social Networks

The fast development of event-based social networks (EBSN) provides a convenient platform for recruiting offline participants via online event announcements. Given its ever-increasing new events, how to accurately recommend users their most preferred ones is a key to the success of an EBSN. In this paper, we propose a semantic-enhanced and context-aware hybrid collaborative filtering for event recommendation, which combines semantic content analysis and contextual event influence for user neighborhood selection. In particular, we first exploit the latent topic model for analyzing event description text and establish each user a long-term interest model and short-term interest model from her event registration history. We next establish each event an influence weight to jointly represent its social impact among users and its semantic uniqueness among events. For one user, we select her neighbors according to their long-term interest similarities weighted by events’ influences. For new event recommendation, we construct a user-event rating matrix based on users’ short-term interest models and for each user, we compute event rating predictions from her neighbors’ ratings. The experiments based on the real-world dataset demonstrate the superiority of our algorithm over the peer schemes.

A Multi-Element Hybrid Location Recommendation Algorithm for Location Based Social Networks

In the environment of data explosion, how to make an effective and accurate personalized point of interest (POI) recommendation in location-based social networks (LBSNs) is a challenging and meaningful task. Fortunately, there is a lot of information that we can use. We can make recommendations by mining the rich information hidden in user check-in records. In this paper, we propose a recommend system named GFP-LORE. Specifically, we have designed a framework, which integrates various influencing factors. First, we modeled friend sign-in frequencies and POI popularity as a power-law distribution and the experiment proved that it is effective. Then, we got the influence of geographic information by modeling it according to the longitude and latitude of the user’s check-in location. After that, we sorted the historical check-in records of all users according to time and then mine an overall pattern of location transfer-order pattern. Then, we combine it with each user’s own unique location transfer record to get the possibility of the user going to the next POI. Finally, we synthesize the above four influence factors into a unified correlation probability rating and recommend a new location by this probability rating. We tested our system on the open real check-in data set, and the results of our simulation experiments show that the recommendation effect of our system is better than those algorithms used in the contrast test.

Sentiment Analysis of Weibo Comment Texts Based on Extended Vocabulary and Convolutional Neural Network

In the era of big data and Internet, social network platforms, blogs, and recommender systems generate thousands of subjective information every day. The emotional content of these information may be related to books, characters, commodities, activities and so on. Analyzing and mining subjective emotional information is conducive to personal decision making, enterprise reform, and government’s public opinion regulation.In this paper, based on Weibo comment texts, we use the network term and the wiki Chinese data set to expand the original vocabulary, train word embeddings and realize the sentence-level sentiment classification based on the convolution neural network. At the same time, an optimization method according to the statement length of pooling layer is put forward. The method is proved to be effective with high accuracy on our data set.

Time and Location Aware Points of Interest Recommendation in Location-Based Social Networks

The wide spread of location-based social networks brings about a huge volume of user check-in data, which facilitates the recommendation of points of interest (POIs). Recent advances on distributed representation shed light on learning low dimensional dense vectors to alleviate the data sparsity problem. Current studies on representation learning for POI recommendation embed both users and POIs in a common latent space, and users’ preference is inferred based on the distance/similarity between a user and a POI. Such an approach is not in accordance with the semantics of users and POIs as they are inherently different objects. In this paper, we present a novel translation-based, time and location aware (TransTL) representation, which models the spatial and temporal information as a relationship connecting users and POIs. Our model generalizes the recent advances in knowledge graph embedding. The basic idea is that the embedding of a pair corresponds to a translation from embeddings of users to POIs. Since the POI embedding should be close to the user embedding plus the relationship vector, the recommendation can be performed by selecting the top-k POIs similar to the translated POI, which are all of the same type of objects. We conduct extensive experiments on two real-world datasets. The results demonstrate that our TransTL model achieves the state-of-the-art performance. It is also much more robust to data sparsity than the baselines.

A self-adaptive point-of-interest recommendation algorithm based on a multi-order Markov model

As one of the personalization technologies, point-of-interest (POI) recommendation systems have attracted more and more attention from academic and industrial researchers. Exploiting the spatio-temporal pattern of users check-ins for user modeling is the core content of the current research of POI recommendation in location-based social networks (LBSNs). In this paper, we propose a POI recommendation algorithm based on a multi-order Markov model, which predicts users next favorite POIs based not only on their current location but also on their previous location, and propose a self-adaptive algorithm to adjust our multi-order Markov model to be available to all users check-ins. Moreover, to improve the precision of our proposed POI recommendation algorithm, we incorporate the geographical influence and temporal popularity of users checked-in POIs into our proposed algorithm. Finally, experimental results on two real datasets demonstrate that our proposed algorithm outperforms the state-of-the-art POI recommendation methods in terms of F−measure@N(N=5,10,15,20).