Finding Top-k Answers in Node Proximity Search using Distribution State Transition Graph

Personalized PageRank (PPR) computation is a fundamental operation in web search, social networks, and graph analysis. Given a graph G , a source s, and a target t, the PPR query Π( s, t ) returns the probability that a random walk on G starting from s terminates at t. Unlike global PageRank which can be effectively pre-computed and materialized, the PPR result depends on both the source and the target, rendering results materialization infeasible for large graphs. Existing indexing techniques have rather limited effectiveness; in fact, the current state-of-the-art solution, BiPPR, answers individual PPR queries without pre-computation or indexing, and yet it outperforms all previous index-based solutions. Motivated by this, we propose HubPPR, an effective indexing scheme for PPR computation with controllable tradeoffs for accuracy, query time, and memory consumption. The main idea is to pre-compute and index auxiliary information for selected hub nodes that are often involved in PPR processing. Going one step further, we extend HubPPR to answer top- k PPR queries, which returns the k nodes with the highest PPR values with respect to a source s , among a given set T of target nodes. Extensive experiments demonstrate that compared to the current best solution BiPPR, HubPPR achieves up to 10x and 220x speedup for PPR and top- k PPR processing, respectively, with moderate memory consumption. Notably, with a single commodity server, HubPPR answers a top- k PPR query in seconds on graphs with billions of edges, with high accuracy and strong result quality guarantees.

With the wide spread of movie platforms and the diversity in the quantity and quality of movies shown, it is difficult for Movie audiences to find related information about the movies that they may be interested in. And have contributed to finding a way to provide a list of movies that suits the user’s desire and personalization. In this study, we present an enhancement to the recommendation system by exploiting the personalized PageRank algorithm to find the most personalized candidate list of movies for the target user. PageRank Algorithm is a useful method for the recommendation task. PageRank algorithm is used to improve the representation of movies and users in the graph network in addition to ranking movies for a target user in the recommendation activity. User preferences have been calculated based on movie genres as a part of building the user profile from his history of rating (content analysis), and for each user, this preference was used to produce a dynamic initial rank value for each movie instead of the traditional static initial rank value. In addition to personalizing the target user, a new way has been presented to personalize each user with different weights according to their rating on movies, by supporting the proposed recommender system with new personalization parameters for each user. To evaluate the performance of the proposed recommender system and measure the accuracy of recommendations, precision, and recall metrics have been used. The experimental results show that there is a significant improvement in the recommendations process, which means the hybrid recommender system using personalized PageRank (HRS-PPR) is better than the traditional recommender system using personalized PageRank. The experimental results show that the proposed approach enhances the accuracy of the recommendation regarding precision and recall measures.

Link-based similarity measures play a significant role in many graph based applications. Consequently, measuring node similarity in a graph is a fundamental problem of graph datamining. Personalized pagerank (PPR) and simrank (SR) have emerged as the most popular and influential link-based similarity measures. Recently, a novel link-based similarity measure, penetrating rank (P-Rank), which enriches SR, was proposed. In practice, PPR, SR and P-Rank scores are calculated by iterative methods. As the number of iterations increases so does the overhead of the calculation. The ideal solution is that computing similarity within the minimum number of iterations is sufficient to guarantee a desired accuracy. However, the existing upper bounds are too coarse to be useful in general. Therefore, we focus on designing an accurate and tight upper bounds for PPR, SR, and P-Rank in the paper. Our upper bounds are designed based on the following intuition: the smaller the difference between the two consecutive iteration steps is, the smaller the difference between the theoretical and iterative similarity scores becomes. Furthermore, we demonstrate the effectiveness of our upper bounds in the scenario of top-k similar nodes queries, where our upper bounds helps accelerate the speed of the query. We also run a comprehensive set of experiments on real world data sets to verify the effectiveness and efficiency of our upper bounds.

Network-based recommendation systems leverage the topology of the underlying graph and the current user context to rank objects in the database. Random walk-based techniques, such as PageRank, encode the structure of the graph in the form of a transition matrix of a stochastic process from which the significances of the nodes in the graph are inferred. Personalized PageRank (PPR) techniques complement this with a seed node set which serves as the personalization context. In this paper, we note (and experimentally show) that PPR algorithms that do not differentiate among the seed nodes may not properly rank nodes in situations where the seed set is incomplete and/or noisy. To tackle this problem, we propose alternative robust personalized PageRank (RPR) strategies, which are insensitive to noise in the set of seed nodes and in which the rankings are not overly biased towards the seed nodes. In particular, we show that novel teleportation-discounting and seed-set maximal PPR techniques help eliminate harmful bias of individual seed nodes and provide effective seed differentiation to lead to more accurate rankings. We also show that the proposed techniques lead to efficient implementations, where existing approximation algorithms and/or parallel implementations for computing the PPR scores can be easily leveraged. Moreover, the proposed formulations are reuse-promoting in the sense that, it is possible to divide the work relative to individual seed nodes and cache the intermediary results obtained during the computation, and especially in systems with large query throughputs, it may be possible to cluster queries based on the partial overlaps between the seed sets and reduce the overall robust PPR computation costs. Experiment results show that the proposed techniques are efficient and highly effective in improving recommendations and eliminating unwanted bias due to imperfections in the seed set.

Personalized PageRank (PPR) is an important variation of PageRank, which is a widely applied popularity measure for Web search. Unlike the original PageRank, PPR is a node proximity measure that represents the degree of closeness among multiple nodes within a graph. It is also widely applied to diverse domains, such as information retrieval, recommendations, and knowledge discovery, due to its theoretical simplicity and flexibility. However, computing PPR in large graphs using naïve algorithms such as iterative matrix multiplication and matrix inversion is not fast enough for many of these applications. Therefore, devising efficient PPR algorithms has been one of the most important subjects in large-scale graph processing. In this paper, we review the algorithms for efficient PPR computations, organizing them into five categories based on their core ideas. Along with detailed explanations including recent advances and their applications, we provide a multifaceted comparison of the algorithms.

Personalized PageRank (PPR) based measures of node proximity have been shown to be highly effective in many prediction and recommendation applications. The use of personalized PageRank for large graphs, however, is difficult due to its high computation cost. In this paper, we propose a Locality-sensitive, Re-use promoting, approximate personalized PageRank (LR-PPR) algorithm for efficiently computing the PPR values relying on the localities of the given seed nodes on the graph: (a) The LR-PPR algorithm is locality sensitive in the sense that it reduces the computational cost of the PPR computation process by focusing on the local neighborhoods of the seed nodes. (b) LR-PPR is re-use promoting in that instead of performing a monolithic computation for the given seed node set using the entire graph, LR-PPR divides the work into localities of the seeds and caches the intermediary results obtained during the computation. These cached results are then reused for future queries sharing seed nodes. Experiment results for different data sets and under different scenarios show that LR-PPR algorithm is highly-efficient and accurate.

The large volume of data available in location-based social networks (LBSNs) enables Point-of-Interest (POI) recommendation services. On another hand, the heterogeneous information (e.g., user check-in records, geographical features of POIs, social network and user reviews) imposes great challenges on effective POI recommendation. In this paper, we focus on leveraging such rich information in an integrated manner to improve POI recommendation performance. We exploit not only the geographical and social information, but also aspects extracted from user reviews to better model users' preferences. More specifically, to fully utilize various types of information, we construct a novel heterogeneous graph, Aspect-aware Geo-Social influence Graph (AGSG), by fusing various relations among the three types of nodes, i.e., users, POIs and aspects. The personalized POI recommendation task is then transformed as a graph node ranking problem in AGSG. We design a graph-based recommendation algorithm based on both personalized PageRank (PPR) and meta paths, to fully exploit the heterogeneous graph structure as well as to capture the semantic relations among the various nodes. Experiments on three real-world datasets show that our proposed approach outperforms the state-of-art methods.

The random walk process on network data is a widely-used approach for network representation learning. However, we argue that the sampling of node sequences and the subsampling for the Skip-gram’s contexts have two drawbacks. One is less possible to precisely find the most correlated context nodes for every central node with only uniform graph search. The other is not easily controlled due to the expensive cost of hyperparameter tuning. Such two drawbacks lead to higher training cost and lower accuracy due to abundant and irrelevant samples. To solve these problems, we compute the adaptive probability of random walk based on Personalized PageRank (PPR), and propose an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Adaptive SKip-gram</i> (ASK) model without using complicated sampling process and negative sampling. We utilize <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> -most important neighbors for positive samples selection, and attach their corresponding PPR probability into the objective function. Based on benchmark datasets with three citation networks and three social networks, we demonstrate the improvement of our ASK model for network representation learning in tasks of link prediction, node classification, and embedding visualization. The results achieve more effective performance and efficient learning time.

Analysis of large graphs is critical to the ongoing growth of search engines and social networks. One class of queries centers around node affinity, often quantified by random-walk distances between node pairs, including hitting time, commute time, and personalized PageRank (PPR). Despite the potential of these "metrics," they are rarely, if ever, used in practice, largely due to extremely high computational costs. In this paper, we investigate methods to scalably and efficiently compute random-walk distances, by "embedding" graphs and distances into points and distances in geometric coordinate spaces. We show that while existing graph coordinate systems (GCS) can accurately estimate shortest path distances, they produce significant errors when embedding random-walk distances. Based on our observations, we propose a new graph embedding system that explicitly accounts for per-node graph properties that affect random walk. Extensive experiments on a range of graphs show that our new approach can accurately estimate both symmetric and asymmetric random-walk distances. Once a graph is embedded, our system can answer queries between any two nodes in 8 microseconds, orders of magnitude faster than existing methods. Finally, we show that our system produces estimates that can replace ground truth in applications with minimal impact on application output.

Social network contains large amounts of user interests and preferences which may therefore help improve techniques for information retrieval, such as user profile construction and query expansion. By the manual annotation, we find that query intention normally consists of several retrieval behaviors and a certain number of social network behaviors. This paper argues that a user's query intention has a strong correlation with the corresponding social network contents, which focuses on two aspects. First, we analyze the proportion of social network websites in users' retrieval records, and evaluate the degree of relationship between information retrieval behavior and social network behavior from a statistic perspective. Second, a similarity between the two components of query intention, retrieval content and social network content, is calculated employing a topic model to measure the correlation of them from a semantic perspective. The evaluation on a ground-truth data of retrieval records indicates that a long-term retrieval procedure is usually accompanied with accessing social network websites. More importantly, the experimental results show that more than 54% of the content from social network has a strong correlation with users' query intention.

Social online learning is gaining prominence with the emergence of social media and networks. In the presence of recommendation systems, the learner has access to several educational contents and elements proposed according to his interaction, profile, dynamics and other parameters. Our proposal revolves around an intelligent recommender system for generating recommendations in social learning environments, in particular social networks, based on the learner's activities, falling into the category of hybrid recommender systems and integrating the concept of community detection in the generation of recommendations. The proposed global approach is composed of several recommender systems: the correlation and co-occurrence based recommender system aims at measuring the degree of connectivity between learners' activities in order to generate relevant recommendations in the light of learners' implicit feedback, the community detection based recommender system through friendships. This system is an important part of our overall approach as it combines the hybrid recommendation system with the community detection based on friendship links between learners and the recommendation system based on the degree of interactivity of learners with the available learning objects. This system also consists in detecting communities, which group learners with similar preferences and orientations. The results obtained prove the potential of combining the two notions of connectivity and the integration of community detection based on several indicators to produce more relevant recommendations, while offering recommendations by group of learners by applying the concept of continuous improvement.KeywordsSocial learningRecommender systemCorrelationCo-occurrenceFriendshipInteractionCommunity detection

A location-based orientation-aware recommender system using IoT smart devices and Social Networks

In this talk, I will overview our recent work addressing the problem of locating useful items and trading partners within large-scale social and agentorganized networks. In agent-organized networks (AONs), agents have a limited number of collaboration partners at any time, represented by edges in a network of agent nodes. Agents have to learn to estimate the utility of current trading partners and rewire edges, i.e., change partners to improve profitability.We show that an exponentially decaying exploration scheme produces similar utilities with much less rewiring costs compared to a previously proposed randompartner selection scheme within a production and exchange economy. We analyze the observed performance and demonstrate the robustness of our proposed rewiring scheme in a more realistic model of the economy that incorporate minimum trade volumes and storage capacities. We have developed a Social Network-based Item Recommendation (SNIR) system consisting of agents that utilizes the connections of a user in the social network to search for items of interest posted by peers. Our approach enables a focused mining of the huge number of items available in the social network to generate targeted recommendations that can improve the precision of the result returned from a user's query. We have implemented the SNIR agent-based framework in Flickr, a photo-sharing social network, for searching and recommendation of photos by using tag lists as search queries. We discuss the architecture of our system, motivate the searching and referral schemes used, and demonstrate the effectiveness of the SNIR approach by presenting representative results.

Recommender system is the system which, by using automated information filtering technology, recommends products or services to the customers who are likely to be interested in. Those systems are widely used in many different Web retailers such as Amazon.com, Netfix.com, and CDNow.com. Various recommender systems have been developed. Among them, Collaborative Filtering (CF) has been known as the most successful and commonly used approach. CF identifies customers whose tastes are similar to those of a given customer, and recommends items those customers have liked in the past. Numerous CF algorithms have been developed to increase the performance of recommender systems. However, the relative performances of CF algorithms are known to be domain and data dependent. It is very time-consuming and expensive to implement and launce a CF recommender system, and also the system unsuited for the given domain provides customers with poor quality recommendations that make them easily annoyed. Therefore, predicting in advance whether the performance of CF recommender system is acceptable or not is practically important and needed. In this study, we propose a decision making guideline which helps decide whether CF is adoptable for a given application with certain transaction data characteristics. Several previous studies reported that sparsity, gray sheep, cold-start, coverage, and serendipity could affect the performance of CF, but the theoretical and empirical justification of such factors is lacking. Recently there are many studies paying attention to Social Network Analysis (SNA) as a method to analyze social relationships among people. SNA is a method to measure and visualize the linkage structure and status focusing on interaction among objects within communication group. CF analyzes the similarity among previous ratings or purchases of each customer, finds the relationships among the customers who have similarities, and then uses the relationships for recommendations. Thus CF can be modeled as a social network in which customers are nodes and purchase relationships between customers are links. Under the assumption that SNA could facilitate an exploration of the topological properties of the network structure that are implicit in transaction data for CF recommendations, we focus on density, clustering coefficient, and centralization which are ones of the most commonly used measures to capture topological properties of the social network structure. While network density, expressed as a proportion of the maximum possible number of links, captures the density of the whole network, the clustering coefficient captures the degree to which the overall network contains localized pockets of dense connectivity. Centralization reflects the extent to which connections are concentrated in a small number of nodes rather than distributed equally among all nodes. We explore how these SNA measures affect the performance of CF performance and how they interact to each other. Our experiments used sales transaction data from H department store, one of the well?known department stores in Korea. Total 396 data set were sampled to construct various types of social networks. The dependant variable measuring process consists of three steps; analysis of customer similarities, construction of a social network, and analysis of social network patterns. We used UCINET 6.0 for SNA. The experiments conducted the 3-way ANOVA which employs three SNA measures as dependant variables, and the recommendation accuracy measured by F1-measure as an independent variable. The experiments report that 1) each of three SNA measures affects the recommendation accuracy, 2) the density's effect to the performance overrides those of clustering coefficient and centralization (i.e., CF adoption is not a good decision if the density is low), and 3) however though the density is low, the performance of CF is comparatively good when the clustering coefficient is low. We expect that these experiment results help firms decide whether CF recommender system is adoptable for their business domain with certain transaction data characteristics.

Driven by many real applications, we study the problem of seeded graph matching. Given two graphs and , and a small set of pre-matched node pairs where and , the problem is to identify a matching between and growing from , such that each pair in the matching corresponds to the same underlying entity. Recent studies on efficient and effective seeded graph matching have drawn a great deal of attention and many popular methods are largely based on exploring the similarity between local structures to identify matching pairs. While these recent techniques work provably well on random graphs, their accuracy is low over many real networks. In this work, we propose to utilize higher-order neighboring information to improve the matching accuracy and efficiency. As a result, a new framework of seeded graph matching is proposed, which employs Personalized PageRank (PPR) to quantify the matching score of each node pair. To further boost the matching accuracy, we propose a novel postponing strategy, which postpones the selection of pairs that have competitors with similar matching scores. We show that the postpone strategy indeed significantly improves the matching accuracy. To improve the scalability of matching large graphs, we also propose efficient approximation techniques based on algorithms for computing PPR heavy hitters. Our comprehensive experimental studies on large-scale real datasets demonstrate that, compared with state-of-the-art approaches, our framework not only increases the precision and recall both by a significant margin but also achieves speed-up up to more than one order of magnitude.